Hibernating mammals can survive several months without feeding by limiting their carbohydrate catabolism and using triacylglycerols stored in white adipose tissue (WAT) as their primary source of fuel. Here we show that a lipolytic enzyme normally found in the gut, pancreatic triacylglycerol lipase (PTL), is expressed in WAT of hibernating 13-lined ground squirrels (Spermophilus tridecemlineatus). PTL expressed in WAT is encoded by an unusual chimeric retroviral-PTL mRNA ∼500 bases longer than the predominant PTL message found in other ground squirrel tissues. Seasonal measurements detect the chimeric mRNA and PTL enzymatic activity in WAT before and during hibernation, with both showing their lowest observed levels 1 wk after hibernation concludes in mid-March. PTL is expressed in addition to hormone-sensitive lipase, the enzyme typically responsible for hydrolysis of triacylglycerols in WAT. Because of the distinct catalytic and regulatory properties of both enzymes, this dual-triacylglycerol lipase system provides a means by which the fuel requirements of hibernating 13-lined ground squirrels can be met without interruption.
- Spermophilus tridecemlineatus
- white adipose tissue
- pancreatic triacylglycerol lipase
during winter, limited supplies of food make energy management an important issue for mammals trying to survive cold, harsh conditions where heat must be continuously generated to replace heat lost to the environment. Consequently, the colder the surroundings, the more energy required for an animal to maintain this equilibrium. Some mammals, however, have the ability to reduce their metabolic rate and survive for up to 6 mo without food in an inactive state where body temperatures can approach 0°C. This inactive state is called hibernation.
Mammals prepare for hibernation during late summer and early fall by depositing lipid in the form of triacylglycerols in their main fat storage depot, the white adipose tissue (WAT) (4, 6). Hibernating mammals have a respiratory quotient of 0.7, indicating that stored lipids, not carbohydrates, are their main source of energy during hibernation (reviewed in Ref. 19). Although fat stores provide the necessary metabolic fuel, mechanisms of lipid mobilization and utilization during hibernation are not completely understood. In euthermic mammals, the enzyme typically responsible for the hydrolysis of triacylglycerols stored in WAT is hormone-sensitive lipase (HSL) (26). The release of nonesterified fatty acids used as fuel in hibernating mammals is also catalyzed by HSL in hibernating and nonhibernating states (8, 20). We present evidence that a second lipolytic enzyme, pancreatic triacylglycerol lipase (PTL), is expressed in the WAT of 13-lined ground squirrels (Spermophilus tridecemlineatus) before and during hibernation.
Until recently, PTL was believed to be expressed exclusively in the pancreatic acinar cells and secreted into the small intestine as a means to digest dietary fat (reviewed in Ref. 17). However, we found that PTL is differentially expressed in the heart of 13-lined ground squirrels, where it provides low-temperature lipolysis during hibernation (2). Here we report that PTL is expressed in WAT in the form of a novel chimeric retroviral-PTL mRNA that is distinct from the PTL message found in the heart. The expression of genes for PTL and HSL is described, as well as PTL enzymatic activity in WAT. We suggest that the unusual expression of PTL in WAT may have arisen from a retroviral insertion event and that PTL plays an important role in the mobilization of fatty acids from stored triacylglycerols during the near-freezing body temperatures associated with hibernation.
Animal care and use were in accordance with Institutional Animal Care and Use Committee guidelines. Thirteen-lined ground squirrels were obtained from TLS Research (Bartlett, IL) and maintained in captivity on a diet of standard rodent chow and sunflower seeds ad libitum. Animals were kept at ambient temperatures of 23°C in August, 17°C in September, 11°C in October, and 5°C from November through mid-March. Animals were maintained in a 12:12-h dark-light cycle from August through the end of October. From November to mid-March, the animals were housed in total darkness with only water ad libitum; under these conditions, all the animals hibernated. Squirrels were observed daily, and hibernation patterns were determined using the sawdust technique (23). Hibernating animals were at least inday 2 of a torpor bout when they were killed, and animals in interbout arousal were collected after at least three previous hibernation bouts of ≥8 days. The hibernation phase of the study was concluded in mid-March by increasing room temperature to 23°C, reestablishing the 12:12-h dark-light cycle, and providing a diet of standard rodent chow and sunflower seeds ad libitum. These active animals were killed and tissues were removed 7 days after cessation of hibernation. Active and hibernating animals were killed by decapitation to exclude potential hibernation-mimicking effects of drugs used for purposes of anesthesia and/or euthanasia. Body temperatures (rectal) were measured at the time of death. Ground squirrel tissue was promptly removed, placed in cryovials, and frozen immediately in liquid nitrogen. Pancreatic tissue was immersed directly in ice-cold guanidinium isothiocyanate, and RNA was prepared immediately because of high levels of RNase activity in the pancreas.
cDNA library construction.
Total RNA was isolated from WAT as described in RNA isolation and Northern blot analysis. Poly(A)+ mRNA was prepared using mini-oligo(dT) cellulose spin columns (5 Prime → 3 Prime) according to manufacturer's instructions. A directional cDNA library was made from WAT mRNA by a modified method of Okayama et al. (21) using the Superscript Plasmid System (GIBCO-BRL Life Technologies). First-strand cDNA was synthesized by Superscript II reverse transcriptase using an oligo(dT) primer containing aNotI site. Second-strand synthesis was performed withEscherichia coli DNA polymerase in the presence of E. coli DNA ligase and RNase H. Adapters containing a SalI overhang were blunt-end ligated onto the double-stranded cDNA with T4 DNA ligase and then digested with NotI. The cDNA fragments with a SalI site at the 5′ end and a NotI site at the 3′ end were ligated into the pSPORT1 high-copy plasmid with T4 DNA ligase and electroporated into ElectroMAX DH10B bacterial cells (GIBCO-BRL Life Technologies). The WAT cDNA library was probed with heart PTL and HSL cDNA fragments. Positive WAT cDNA clones were sequenced by automated cycle sequencing using fluorescently tagged terminator base analogs (Applied Biosystems). Sequence analysis was performed using MacVector and AssemblyLIGN software (Oxford Molecular Group).
RNA isolation and Northern blot analysis.
Total RNA was prepared from ground squirrel tissues by a modification of the method developed by Chomczynski and Sacchi (5). Tissues were homogenized with a rotary-blade homogenizer in 4 M guanidinium isothiocyanate, and sodium acetate was added to a final concentration of 0.2 M. The WAT homogenate was subjected to an initial centrifugation for 10 min at 3,000 g in a swinging-bucket rotor, then the glycerol layer and lipid cake were removed. The homogenate for all tissues was extracted once with water-saturated phenol and chloroform-isoamyl alcohol (49:1) and then extracted twice with phenol-chloroform-isoamyl alcohol (50:49:1). All extractions were centrifuged in a swinging-bucket rotor at 4,500 g through Phase Lock Gel (5 Prime → 3 Prime). RNA was precipitated with isopropanol, rinsed with ethanol, dissolved in water, and quantified by absorption spectrophotometry.
Total RNA from ground squirrel tissues was separated on 1.2% agarose gels containing formaldehyde and then transferred onto Magnacharge nylon membrane (MSI) using 10× sodium chloride-sodium phosphate-EDTA according to the method of Sambrook et al. (25). Each blot contained 8 μg of total WAT RNA from animals at several points during the hibernation season or 10 μg of total RNA from various tissues and organs to show the distribution of PTL mRNA in hibernating ground squirrels. Northern blots were hybridized with 32P-labeled PTL and HSL cDNAs. All blots were also hybridized with a cDNA encoding human β-actin (11) as a control for gel loading and as an indicator of RNA integrity. PTL and HSL mRNA levels were analyzed by autoradiography and quantified using a STORM PhosphorImager and ImageQUANT software (Molecular Dynamics). Amounts of PTL and HSL mRNAs were normalized to β-actin within the same gel lane, and levels of upregulation were calculated relative to euthermic animals taken in the spring. One-way ANOVA was used to determine whether there was a significant difference in PTL and HSL mean mRNA levels relative to their spring counterparts. All statistical calculations were performed using Microsoft Excel software.
Preparation of WAT acetone powders and PTL enzyme assays.
WAT protein extracts containing active PTL were delipidated by the preparation of acetone powder from WAT (2, 10). The acetone powder was solubilized in 0.2 M Tris · HCl (pH 9.2) and 1 M ethylene glycol and reconstituted by gentle agitation for 2–3 h at 4°C. The total solubilized protein concentration was determined by the Bradford assay using dye reagents from Bio-Rad Laboratories. PTL enzymatic activity of solubilized acetone powders was analyzed by a modification of the methods described by Lowe (14, 15) using 14C-labeled triolein (NEN) as a substrate. Assays were performed in the presence of the inhibitory bile salt sodium deoxycholate (19.8 mM). PTL activity was recovered by the addition of 2.8 μg/ml of the PTL-specific activator protein colipase (Sigma). The substrate was emulsified by sonication and added to soluble WAT protein extracts for a final reaction volume of 60 μl. Reactions were incubated at 37°C for 10 min and terminated with the addition of 975 μl of methanol-chloroform-heptane (14:12.5:10 vol/vol/vol) and 315 μl of 0.1 M boric acid and 0.1 M K2CO3 (pH 10.5). After organic extraction, 100 μl of the aqueous layer were counted in 5 ml of scintillation fluid. PTL activity was calculated on the basis of nanomolar free fatty acids liberated per minute per milligram of total soluble WAT protein present in the resuspension. One-way ANOVA was used to determine whether there were significant differences in PTL activity of the WAT extracts in relation to their spring (March) counterparts.
Northern blots containing RNA prepared from 10 different tissues from hibernating 13-lined ground squirrels were probed with cDNAs encoding HSL and PTL. Analysis of 10 μg of total RNA from each tissue shows that HSL is seen primarily in WAT, with much weaker signals detected in heart, skeletal muscle, and testes (Fig.1 A). Hybridization of a different multitissue blot (Fig. 1 B) with a probe for ground squirrel heart PTL (accession no. AF027293) shows a 1.8-kb PTL mRNA in heart, pancreas, and testes. Unexpectedly, a 2.3-kb mRNA in WAT also hybridizes to the PTL probe. Because fat utilization during hibernation is critical to animal survival, and WAT is the main fat storage depot, we decided to investigate the expression of these two triacylglycerol lipases.
Full-length clones corresponding to HSL and PTL were isolated from a hibernating WAT cDNA library and subsequently characterized. The cDNA for ground squirrel HSL (accession no. AF177401) is 2.8 kb long and contains an open reading frame encoding a 763-amino acid protein that shares 87% amino acid identity with human HSL (13). In the case of PTL expressed in WAT of 13-lined ground squirrels, two different cDNA clones were identified that corresponded with the ∼2.3-kb size of the PTL transcript seen on RNA blots (Fig. 1). Nucleotide sequence analysis reveals that both PTL clones, denoted 7G5 and 22A4 (accession nos. AF177402 and AF177403, respectively), contain the entire PTL open reading frame, showing 99.8% nucleotide identity with the protein coding region of the ground squirrel heart PTL cDNA (2). The predicted 465-amino acid protein is identical to heart PTL and shares 85% amino acid identity with PTL expressed in the human pancreas (18). The 3′-untranslated regions (3′-UTR) of the heart and WAT clones are 98.2% identical at the nucleotide level.
The greater size of the two WAT clones compared with the heart PTL cDNA arises from a distinct 5′-untranslated region (5′-UTR) beginning three bases upstream of the PTL start codon (Fig.2 A). Unlike the heart PTL cDNA, the 5′-UTRs of 7G5 and 22A4 consist predominantly of sequences homologous to mammalian retroviral genomes based on GenBank comparisons. Both WAT cDNAs show the highest homology with an active porcine endogenous retrovirus (1). Segments of retroviralgag and pol genes comprise the majority of the 5′ regions in the two different WAT PTL cDNAs shown in Fig. 2 B. Clone 22A4 contains 542 nucleotides of retroviral-like sequence; clone 7G5 shows a 706-nucleotide homology that includes the start codon of the gag gene. Neither cDNA clone contains sequence homologous to the env gene. The linear arrangement of these genes in retroviral genomes is conserved in the gene segments found in the two WAT cDNAs, thus resembling spliced exons from a larger chimeric retroviral-PTL gene.
The retroviral region of both WAT cDNAs abuts the PTL open reading frame at the same point within the sequence encoding the viral integrase protein of the pol gene; however, the integrase open reading frame is not in frame with the PTL start codon. It is surprising that the nucleotide sequences of the two WAT cDNAs and the heart cDNA are nearly identical in the PTL coding region and 3′-UTR yet so markedly different immediately upstream of the start codon. The different-sized transcripts (2.3 vs. 1.8 kb) and the divergent 5′-UTRs of the WAT and heart mRNAs suggest a single PTL gene with alternative splicing of the PTL mRNA, or two separate genes. It is possible, as will be discussed later, that proximal insertion of retroviral DNA has conferred beneficial expression of PTL in the WAT of hibernating 13-lined ground squirrels.
To investigate the expression of PTL and HSL during hibernation, total WAT RNA was prepared from individual animals at various states of activity from August through March (Fig.3 A). Similar to heart PTL expression seen in active animals during the fall (2), levels of PTL mRNA in WAT are highest in October before hibernation. During hibernation, mRNA levels remain fairly constant regardless of body temperature and the state of animal activity (Fig. 3 B). This includes regularly occurring interbout arousals, where body temperatures can approach 37°C for up to 24 h (29). However, 1 wk after the final arousal in March, PTL mRNA is barely detectable via Northern blot analysis. When normalized to β-actin mRNA on individual blots, the relative level of PTL mRNA from August through December is elevated from 2.5 to 4.5 times greater than that seen in the active March animals. On the other hand, HSL mRNA levels remain fairly constant, showing nonsignificant differences in expression from August to March (Fig. 3 B). This observation contrasts with the previously reported upregulation of HSL in the WAT of marmots (30), suggesting species-specific regulation of HSL during hibernation.
An in vitro lipase assay (15) was used to investigate PTL activity in WAT protein extracts (10) prepared from animals collected throughout the hibernation season. Because WAT contains other triacylglycerol lipases (HSL and lipoprotein lipase), it was necessary to include the inhibitory bile salt sodium deoxycholate in every reaction. Inhibition of lipolysis is overcome by the addition of the PTL-specific activator protein colipase, which selectively stabilizes the tertiary structure of PTL in the presence of bile salts (16). As a control for possible PTL contamination of the colipase, the colipase mixture was incubated with soluble WAT protein that had been heated at 100°C for 5 min. Neither colipase with the heat-treated WAT extract, nor colipase alone, showed any PTL activity under assay conditions (data not shown). Colipase-dependent PTL activity was observed in all WAT samples assayed at 37°C (Fig.4). As with the relative levels of PTL mRNA (Fig. 3 B), PTL-mediated lipolysis declines to its lowest level 1 wk after activity resumes in March; however, none of the changes seen with PTL activity were determined to be significant.
This report provides the first evidence, in any animal, that pancreatic triacylglycerol lipase is expressed in WAT. Our finding that PTL is expressed in WAT of 13-lined ground squirrels indicates a new mechanism by which a hibernating mammal can mobilize fatty acids from its main fat storage depot. The fact that mRNAs encoding PTL in WAT contain large tracts of retroviral sequence at their 5′ end raises the additional question of whether this novel expression is the result of a retroviral insertion. The insertion of retroviral DNA into a host genome has been documented as a vehicle for evolutionary change (reviewed in Ref. 24). As insertion elements, retroviruses can downregulate normal patterns of gene expression by disrupting the coding regions and promoter elements of various genes. However, in some instances, insertion of retroviral sequence can upregulate host gene expression by providing an enhancer element that activates a gene in a tissue where it is normally inactive (27).
Figure 5 shows a model where insertion of retroviral DNA near the 5′ end of a duplicated PTL gene providescis-acting sequences that selectively activate PTL in an unusual location such as WAT. Precedent for retroviral-mediated activation has been reported with another digestive enzyme of pancreatic origin, human salivary amylase (27). Sequences derived from the insertion of a retrovirus ∼40 million years ago are responsible for parotid gland-specific expression of the salivary amylase genes. Three salivary amylase genes expressed in the parotid gland are each associated with an intact retroviral element, whereas the two pancreatic amylase genes (not expressed in the parotid) lack retroviral sequence or contain a single long terminal repeat as the remnant of a possible retroviral excision (27). We plan to investigate the genomic organization of the PTL gene(s) in 13-lined ground squirrels to determine the juxtaposition of retroviral and PTL sequences that give rise to the chimeric mRNA seen in WAT. This analysis will also determine whether the ground squirrel PTL gene is duplicated or whether the chimeric mRNA is the result of an alternative splice product from a single PTL gene.
We previously showed that extracts containing PTL from the heart of hibernating 13-lined ground squirrels showed 61% and 34% maximal colipase-dependent activity at 7°C and 0°C, respectively (2). In addition, HSL from rats has also been reported to work as a cold-adapted lipase (13). If low-temperature lipolysis is a general property of HSL, the following question can then be asked: What advantage is conferred by the expression of PTL in WAT? Here we present three considerations as to why the expression of PTL, vs. HSL alone, has a positive influence on 13-lined ground squirrels during hibernation. 1) HSL activity is dependent on specific hormone levels. Catecholamines stimulate HSL activity via interaction with β-adrenergic receptors (reviewed in Ref. 12), while insulin exerts an antilipolytic effect by inhibiting HSL (reviewed in Ref. 7). Alternatively, the activity of PTL is hormone independent and, therefore, not subject to changes in metabolic hormone levels that occur during the hibernation season (3, 20).2) The energy requirements of HSL activation are greater than those of PTL. The signal transduction pathway leading to HSL activation requires ATP for cAMP formation and HSL phosphorylation; PTL activity does not require these ATP-dependent reactions. 3) PTL shows far better catalytic activity than HSL. Tissue-purified and recombinant HSL have activities of ∼220 μmol fatty acid released · min−1 · mg protein−1 at 37°C with an 18-carbon-chain diacylglycerol substrate (22) and shows 10-fold less activity with equivalent-length triacylglycerol substrates (9). Activities of tissue-purified and recombinant PTL range from 1,650 to 2,000 μmol fatty acid released · min−1 · mg protein−1 at 37°C with the 18-carbon-chain triacylglycerol substrate triolein (31), therefore showing a ≥75-fold activity than HSL.
We propose that expression of PTL in WAT may have been initiated by a retroviral insertion event that occurred during the evolution of 13-lined ground squirrels. This novel expression has been selected and maintained, because it provides a robust, alternative triacylglycerol lipase that can remain active during periods of reduced HSL activity. These periods may include late summer and early fall, when the animal is preparing to hibernate and insulin levels are relatively high (28). During hibernation, PTL could also serve as a supplement to HSL when the animal is not feeding and the utilization of fat stores is critical for survival. Considering the physiological extremes of long-term starvation and near-freezing torpor, the expression of PTL and HSL in WAT provides a mechanism where one enzyme could substitute for the other during hibernation. This dual-triacylglycerol lipase system would ensure that the fuel requirements of hibernating 13-lined ground squirrels could be met without interruption.
A growing number of advances in genomic and bioinformatic technologies have recently allowed systematic investigations of previously intractable biological systems. Understudied species showing unique physiological properties are now on the brink of being explored at a level of detail only previously realized with certain model organisms. Mammalian hibernation is one example of a complex physiological process that is now poised for such an investigation. As illustrated in this report, researchers have begun to characterize the genes responsible for the physiological characteristics of hibernation with the intent of escalating this effort through the use of high-throughput genomic and proteomic strategies. The potentially wide functional temperature range of proteins expressed during hibernation is especially attractive because of various practical uses and versatility of handling. Another possible application derived from the strategies of mammalian hibernation includes the achievement of stasis states for purposes of enhanced organ preservation. Identifying the gene products that specify the physiological extremes of the hibernating phenotype has the potential of greatly increasing our understanding of responses to human stresses such as hypoxia, hypothermia, and starvation.
We thank L. B. Mamo and M. B. Rollins for technical assistance and D. G. Cookmeyer, R. Swanstrom, and M. M. Tredrea for critical review of the manuscript.
This work was supported by US Army Research Office Grant DAAD19-01-1-0014 and Augmentation Awards for Science and Engineering Training DAAG55-97-1-0175 and by North Carolina Biotechnology Center Grant 9805-ARG-0038.
Address for reprint requests and other correspondence: M. T. Andrews, Dept. of Biochemistry and Molecular Biology, University of Minnesota School of Medicine, 10 University Dr., Duluth, MN 55812 (E-mail:).
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.
- Copyright © 2001 the American Physiological Society