Proteomic analysis of the winter-protected phenotype of hibernating ground squirrel intestine

Sandra L. Martin, L. Elaine Epperson, James C. Rose, Courtney C. Kurtz, Cécile Ané, Hannah V. Carey


The intestine of hibernating ground squirrels is protected against damage by ischemia-reperfusion (I/R) injury. This resistance does not depend on the low body temperature of torpor; rather, it is exhibited during natural interbout arousals that periodically return hibernating animals to euthermia. Here we use fluorescence two-dimensional difference gel electrophoresis (DIGE) to identify protein spot differences in intestines of 13-lined ground squirrels in the sensitive and protected phases of the circannual hibernation cycle, comparing sham-treated control animals with those exposed to I/R. Protein spot differences distinguished the sham-treated summer and hibernating samples, as well as the response to I/R between summer and hibernating intestines. The majority of protein changes among these groups were attributed to a seasonal difference between summer and winter hibernators. Many of the protein spots that differed were unambiguously identified by high-pressure liquid chromatography followed by tandem mass spectrometry of their constituent peptides. Western blot analysis confirmed significant upregulation for three of the proteins, albumin, apolipoprotein A-I, and ubiquitin hydrolase L1, that were identified in the DIGE analysis as increased in sham-treated hibernating squirrels compared with sham-treated summer squirrels. This study identifies several candidate proteins that may contribute to hibernation-induced protection of the gut during natural torpor-arousal cycles and experimental I/R injury. It also reveals the importance of enterocyte maturation in defining the hibernating gut proteome and the role of changing cell populations for the differences between sham and I/R-treated summer animals.

  • 13-lined ground squirrel
  • ischemia-reperfusion
  • mucosa
  • Spermophilus tridecemlineatus

hibernating mammals switch into a heterothermic phenotype each winter (reviewed in Ref. 8). This seasonal heterothermy is composed of numerous multiday bouts of torpor, with core body temperatures (Tb) reaching as low as −2.9°C (5). Periods of torpor are punctuated by returns to more standard mammalian Tb values of ∼37°C, known as interbout arousal (IBA). Rewarming occurs even under continuous cold environmental temperature, where it necessarily relies solely on endogenous mechanisms of heat production. In addition to Tb, the torpid state in hibernation is characterized by dramatic reductions in metabolic, heart, and respiratory rates. It appears that hibernators exploit the cold to save the energy that would be required to remain homeothermic in winter when food is scarce (reviewed in Refs. 6, 8, 37).

In addition to the energy-saving state of torpor, the hibernator's extraordinary winter physiological phenotype is accompanied by enhanced protection against a number of injury-provoking challenges to mammalian homeostasis. The natural ability of hibernating ground squirrels to survive core Tb values close to and below 0°C is itself an indisputable example of protection, as such temperatures are inevitably lethal in nonhibernating mammals. Hibernators do not show classic signs of cell and tissue damage as they cycle through torpor and arousal, despite evidence that they do experience reduced oxygen delivery, particularly during arousal (16, 23, 35). Several organ systems in hibernators are also notably resistant to a variety of experimental challenges to homeostasis. For example, hippocampal slices from the brains of hibernating ground squirrels are more resistant to ischemia-reperfusion (I/R) injury than those taken from rats or nonhibernating squirrels (15). Hippocampus and striatum are also protected in vivo following a simulated cardiac arrest in arctic ground squirrels (11). Livers from animals in the winter-heterothermic phase of their circannual cycle, whether taken from euthermic ground squirrels during IBA or from torpid animals, are resistant to the damage associated with reperfusion after prolonged cold storage (22). Finally, central to this report is the observation that intestine of IBA hibernators is also resistant to I/R injury.

Using two experimental models of I/R injury, Kurtz et al. (20) demonstrated that intestine from IBA hibernators exhibits significantly less morphological damage when exposed to an I/R insult than intestine from summer ground squirrels or rats. In contrast to the IBA hibernators, both summer ground squirrels and rats showed significant loss of villus epithelium, loss of overall villus structure, infiltration of immune cells (including neutrophils, as measured by elevated myeloperoxidase activity), and congestion of blood vessels following the I/R treatment. In addition, intestinal apoptosis was significantly elevated following I/R over the untreated sham group in both rat and summer ground squirrels, but not in IBA hibernators. These data provide compelling evidence that it is the hibernation season, and not simply the low Tb of torpor, that confers a protected phenotype to the ground squirrel intestine.

To identify molecular components potentially involved in the protected winter phenotype of these hibernators, we applied a powerful two-dimensional (2D) gel approach to quantify changes in a subset of the intestinal proteome in four treatment groups: summer and IBA hibernators subjected to either I/R or sham surgery. Those protein spots that differed significantly among any of these groups were then identified using mass spectrometry. As expected, protein spot differences were found between the summer and IBA hibernator sham-treated intestines, consistent with a significant change in gene expression patterns between the summer and hibernating states of the circannual rhythm (3, 13, 17, 31, 42, 43). There were also protein spot differences between sham and I/R-treated intestines of summer squirrels, and between intestines from sham and I/R-treated IBA hibernators. This latter result reflects the substantial damage induced by I/R in the summer animals compared with the minimal damage in the hibernating animals (20).


Animals and tissue.

All animal use was approved by the University of Wisconsin Institutional Animal Care and Use Committee. Summer squirrels were housed individually with free access to water and food (Purina rodent chow 5001, supplemented with sunflower seeds) at 22°C with a 12:12-h light-dark cycle. In September, squirrels were transferred to a cold room maintained at 4°C. Food and water were removed after squirrels began regular bouts of torpor. Euthermic squirrels were used in summer or IBA during the hibernation season. Specific details of the animals and the I/R protocol used are described elsewhere (20). In brief, after induction of anesthesia with isoflurane, the abdominal cavity was opened, and the superior mesenteric artery was occluded with a vascular clamp for 20 min. The clamp was then removed, and animals were euthanized after a 60-min reperfusion period. These time points were chosen based on pilot studies that showed they induced significant mucosal damage in rats and summer ground squirrels without leading to mortality (20). The entire small intestine was removed, and a 1-cm segment was fixed in 10% formalin for histological analysis. A 3- to 5-cm segment of intact intestine (including mucosa, submucosa, and muscle layers) was frozen in N2 (liquid), shipped on dry ice, and stored at −80°C until homogenization. Sham surgeries were identical to I/R surgeries, but the superior mesenteric artery was not clamped.

Protein extracts.

Approximately 100 mg of tissue were homogenized using a Polytron (Brinkmann Instruments) in buffer containing 0.5 M sucrose, 100 mM phosphate, pH 6.7, 5 mM MgCl2, 1 mM phenylmethylsulfonyl fluoride, and 10 μg/ml protease inhibitors (P8340 Protease Inhibitor Cocktail, Sigma Chemical). The homogenate was transferred to a microfuge tube, and nuclei were pelleted at 500 g, 4°C, for 10 min. The postnuclear supernatant was recovered into a clean tube, divided into 15-μl aliquots, and frozen in N2 (liquid). Aliquots were stored at −80°C; each aliquot was thawed and used only once. Protein concentration was determined by bicinchoninic acid assay (Pierce).

Difference gel electrophoresis labeling.

Equal amounts (μg) of each protein sample were combined into a single tube and mixed thoroughly to create a reference standard. The reference included 12 samples with 3 each of sham and I/R treated from both summer and IBA animals. These mixtures were aliquoted, frozen in N2 (liquid), and stored at −80°C. For labeling, 90 μg of each sample were denatured overnight at room temperature in denaturing buffer: 8 M urea, 2 M thiourea, 4% CHAPS, 25 mM Tris, pH 8.8. Denatured proteins were labeled the next day with either Cy2, Cy3, or Cy5 (CyDye DIGE Fluors, GE Healthcare, Piscataway, NJ). Cy2 was always used to label the reference standard, and Cy3 and Cy5 were alternated, with three samples in each group of six being labeled with Cy3 and the other three with Cy5. Labeling was done according to the manufacturer's protocol, except that the ratio of protein to dye was 50 μg to 80 nmol.

2D gels.

For each gel, a sample labeled with Cy3 and another from a different state labeled with Cy5 were combined with the reference standard (Cy2). The mixed, labeled proteins for each gel were coprecipitated with methanol-chloroform (41); ∼150 μg of total protein were recovered after precipitation. Precipitated proteins were resuspended in 150-μl iso and 150-μl 3 (iso: 9 M urea, 4% CHAPS, 65 mM DTT, 35 mM Tris base, 0.0025% bromophenol blue; 3: 7 M urea, 4% CHAPS, 100 mM DTT, 0.0025% bromophenol blue, 2 M thiourea, 0.8% 3.5–10 ampholytes “Resolyte”; Gallard-Schlesinger, Plainview, NY), and absorbed into Immobiline DryStrips pH 3–10 NL, 18 cm (GE Healthcare). After 18–24 h, the strips were transferred to a Multiphor II isoelectric focusing apparatus (GE Healthcare) and focused with 2 mA and 5 W for all steps: 1) ramp to 500 V over 30 min; 2) ramp to 3,500 V over 4 h; and 3) hold at 3,500 V for 14–17 h. The strips were then incubated 15 min each in reducing (50 mM Tris, pH 6.8, 2% SDS, 15% glycerol, 6 M urea, 1% DTT) and alkylating (50 mM Tris, pH 6.8, 2% SDS, 15% glycerol, 6 M urea, 1.25% iodoacetamide, 0.05% bromophenol blue) buffers before SDS-PAGE (9–16% acrylamide gradient gels). Each gel was scanned in the glass plates with three lasers (Typhoon 9400, GE Healthcare) to collect the Cy2, Cy3, and Cy5 images within 4 h of completion of electrophoresis.

Gels used for spot picking were run as above, except that the starting sample was 270 μg of unlabeled reference standard. The second dimension gel was poured onto a Bind-Silane (PlusOne, GE Healthcare)-treated plate to immobilize the gel for robotic picking. Following electrophoresis, these gels were fixed ≥1 h in 10% methanol, 7.5% acetic acid, stained in SYPRO Ruby Gel Stain (Bio-Rad, Hercules, CA) overnight, and destained in 10% methanol and 7.5% acetic acid before scanning with the green laser.

Quantitative analysis of 2D gels.

For analysis of the four-state comparison among summer and IBA sham and I/R-treated samples, 24 Cy2 gel images and four pick gel images (Sypro stained) were matched. For both sets of data, ∼40 spots were manually matched (“landmarks”) on all of the images in the Biological Variation Analysis module of DeCyder 6.5. The software was then used to automatically match the remaining spots. Cy3 and Cy5 spot values (pixel volumes) were normalized to their corresponding Cy2 spot value, and these normalized values were used for statistical analysis. The data were exported, and analysis of variance was applied to each spot that was found on all 24 gel images, using standard ANOVA, as well as one-way ANOVA, which does not assume equal variances among groups (40). Computations were automated using R (34). To account for multiple testing, q values were obtained to measure false discovery rates, using the “qvalue” R package (33). Spots with q ≤ 0.05 were further analyzed to identify significant pairwise comparisons using Tukey's method. Welch's t-tests, not assuming equal variance, were also performed on those spots with q < 0.05, followed by a Bonferroni correction for the six pairwise comparisons.

The spot numbers with significant q values were examined on all 24 Cy2 images for reproducibility and clear correlation to the pick gel images (Sypro) for picking. Eighty-eight spots were chosen as suitable for attempted protein identification by liquid chromatography followed by tandem mass spectrometry (LC-MS/MS) after this comprehensive examination.

Protein recovery and identification.

Spots were picked into 96-well plates from the Sypro-stained pick gels by an Ettan robotic spot picker (software version 1.10, GE Healthcare) using a 1.5-mm-deep, 1.4-mm-diameter picker head. The plates with picked spot plugs were transferred to a second robot (Ettan Digester, software version 1.10, GE Healthcare) for digestion with trypsin. Picks and digests were done in the University of Colorado Health Sciences Center Proteomics Shared Facility ( Digested gel spots were kept in 96-well plates at −20°C until analysis by LC-MS/MS. Each well contained the acrylamide plug and 5 μl of liquid containing tryptic fragments that eluted from the gel. The liquid containing the peptides was recovered with 10 μl of 5% formic acid into a new 96-well plate. Each sample was applied by HPLC (1100 series pump, Agilent Technologies, Wilmington, DE) onto a C18 reverse-phase trapping column (Zorbax 300SB, 5 μm, Agilent Technologies) for aqueous wash before flowing to a C18 analytic column (Zorbax 300SB, 3.5 μm, Agilent Technologies) and being eluted by nanospray over a 60-min gradient from 0 to 90% buffer B, where buffer A is 99.9% water (HPLC grade, Burdick and Jackson) and 0.1% formic acid, and buffer B is 90% acetonitrile (HPLC grade, Fisher Scientific), 9.9% water (HPLC grade, Burdick and Jackson), and 0.1% formic acid. Full and tandem mass spectra were collected for each spot in an Agilent XCT Plus ion trap on the Ultra Scan setting. A full MS scan was followed by tandem mass spectral (MS/MS) scans of the two highest peaks with a dynamic exclusion of 1 min. Raw mass spectra data were analyzed using Spectrum Mill MS Proteomics Workbench Rev A.03.02.060a ETD-65. The range of mass limits for the precursor ions was 600–4,000 Da, parent and fragment masses were both set to monoisotopic, precursor peptide mass tolerance was 2.5 Da, fragment ion tolerance was 0.7, the enzyme specified was trypsin, maximum number of internal missed cleavage sites was two, and cysteines were given a fixed modification of +57. The database used was an in-house compilation of all National Center for Biotechnology Information (NCBI) mammal sequences in January 2007. It contained 456,753 entries. Most data files were reanalyzed in homology mode using a saved results file, which typically increased protein coverage, often substantially. An in-house program, ExtracTags (available upon request), was written to collate homology peptides onto a single sequence; the data tables report the best score from all of the mapped peptides. If multiple protein IDs were recovered for a single gel spot, only those with at least two peptides and a score >30 are reported. Additional considerations were the congruence between the reported molecular weight and the migration of the gel spot (a database entry within 15% of the observed gel molecular weight), as well as the relative intensity of peptides from the multiple IDs (if the intensity of all of the peptides from one ID was at least 4 times greater than any other). Spectral data from the same spot from up to four pick gels were combined to increase confidence in protein identifications.

The NCBI Entrez gene database ( was used to obtain functional information for the identified proteins. Because human genes generally appear to contain the most detailed information, human gene annotations were used. If the original top-matching protein homolog was not from human, the highest scoring human homolog was identified using BLink.

Western blot analysis.

Proteins (20–40 μg) from intact intestine of sham-treated summer and IBA ground squirrels were separated by SDS-PAGE and transferred to nitrocellulose membranes. Similar protein loading among lanes in each gel was confirmed by Ponceau S staining of membranes before immunoblotting. Antibodies included anti-rat albumin (ALB; no. A110-125, Bethyl Laboratories), anti-human apolipoprotein A-I (APOA1; no. 600-101-109, Rockland), and ubiquitin hydrolase L1 [ubiquitin carboxyl-terminal hydrolase isozyme L1 (UCHL1); PGP9.5, no. 7863-0504, AbD Serotec]. ALB and APOA1 protein bands were detected using the Pierce Chemiluminescence kit. To ensure signals were in the linear range of the X-ray film, immunoblots were exposed for variable lengths of time; all values were corrected for exposure times. Ubiquitin hydrolase was detected after transfer to polyvinylidene difluoride membrane and development with ECL Plus (GE Healthcare). Images were captured with the Typhoon 9400 (GE Healthcare) in the linear range of detection (no film). All bands were quantified using ImageQuant Software (GE Healthcare) and expressed as arbitrary densitometric units. Significant differences in protein expression between sham-treated summer and IBA hibernator groups were determined by t-test.


We hypothesized that the remarkable resistance of the hibernating gut to damage by I/R is due, at least in part, to seasonal changes in the amounts and/or activities of intestinal proteins between summer and hibernating animals. This hypothesis derives from the observation that protection from I/R damage is independent of Tb, because it is apparent in euthermic IBA hibernators in winter (20). Changes in the relative abundance of a protein in two samples can be quantified on 2D gels. The same gels can also reveal some posttranslational modifications (e.g., phosphorylation) that are known to modulate protein activity. To assess changes in the intestinal proteome associated with the winter-protected phenotype of hibernation and the changes associated with the severe damage caused by I/R in summer but not in winter, proteins were isolated from tissues collected from sham and I/R-treated intestines. The 20-min ischemia/60-min reperfusion protocol induces substantial morphological damage in laboratory rats (20) and in summer ground squirrels, but not in the IBA hibernators (Fig. 1; see also Ref. 20). Homogenates of intact intestine from six individuals from each of these four sample groups [summer active sham (SA), aroused hibernator sham (IBA), SA after I/R treatment (SA-IR), and IBA after I/R treatment (IBA-IR)] were analyzed using a four-way difference gel electrophoresis (DIGE) comparison to screen for protein differences. After labeling and fractionation by 2D gel electrophoresis, the reference gel (Fig. 2) in this experiment resolved 2,696 protein spots, which were then compared across all gel images. One thousand twelve spots were found on all 24 gel images; only these were considered for further analysis. The relative intensities of the gel spots were analyzed for significant differences using ANOVA and post hoc pairwise comparisons. Standard ANOVA revealed 120 spots detected in all 24 samples that differed significantly among the animal groups, after controlling for a 5% false discovery rate (q ≤ 0.05). Because some sample groups exhibited unequal variance, spot intensities were also analyzed by one-way ANOVA, allowing unequal variances, which revealed 111 significant spots (q ≤ 0.05). Seventy-eight spots were shared between these two lists. It is noteworthy that only ∼15% of the protein spots (153/1,012) differed significantly among any of the animal groups, which indicates that the abundant, soluble protein component of the intestinal proteome is mostly unchanged, either during hibernation or upon exposure to ischemia followed by reperfusion. All 153 spots from the combined lists were examined for pairwise differences (Tukey's; Table 1) and for suitability for protein identification. The final pick list consisted of 88 protein spots.

Fig. 1.

Effect of ischemia-reperfusion (I/R) treatment on small intestine in summer vs. interbout arousal (IBA) hibernating ground squirrels, as detailed in Ref. 20. Representative photomicrographs are of small intestinal tissues from sham-treated summer squirrel (A), I/R-treated summer squirrel (B), and I/R-treated IBA hibernator (C). The scale bar represents 100 μm; all images are at the same magnification. Only tissue from summer squirrels exhibits a dramatic loss of intestinal villus organization (arrowheads) following I/R treatment (quantified in Ref. 20). Tissues from sham-treated hibernators were similar to the image in C.

Fig. 2.

Two-dimensional (2D) gel fractionation of the intestinal proteome. The numbers on the left of this SYPRO Ruby stained pick gel indicate the positions and approximate molecular mass (in kDa) of marker proteins run on the same gel; numbers above represent the approximate pH gradient across the gel. Solid circles are markers for the automated spot picker. The picked and identified protein spots that differed significantly in this analysis are labeled with either their spot number or Entrez names (for multiple spots containing the same protein); see also Tables 37.

View this table:
Table 1.

Significant pairwise differences for all spots

In the full list of spots, there were 56 protein spot differences between SA and IBA, compared with 43 between SA and SA-IR, and 10 between IBA and IBA-IR (Table 1). These numbers demonstrate a seasonal difference in the intestinal proteome in the sham-treated animals, as well as a larger impact of the I/R treatment in summer compared with winter animals. This basic finding holds for the pickable spots, where there are 35 differences between SA and IBA, 20 between SA and SA-IR, and 9 between IBA and IBA-IR (Table 1). Interestingly, of the 20 pickable protein spots that differ between SA and SA-IR, 15 are increased in intensity after IR treatment and 5 are decreased, whereas, in IBA, only 1 is increased following IR and 8 are decreased (Table 2). This result reveals a difference in the intestinal proteome in SA vs. IBA hibernators, as well as differences in the response to I/R, consistent with the differences in the summer and winter response to I/R that are apparent morphologically.

View this table:
Table 2.

Significant differences of intestinal protein spots and their pairwise fold changes

To identify the specific protein changes revealed by the 2D gel spot analysis, the subset of significantly different protein spots that could be discerned reproducibly on the pick gels were excised, and their tryptic peptides were analyzed by LC-MS/MS. Unambiguous identification was obtained for 73 protein spots (Table 3), 4 spots were ambiguous with 2 proteins identified (Supplementary Table 1), and 11 spots were not identified. (The online version of this article contains supplemental data.) Forty-one of the unambiguously identified spots contained distinct proteins, and 10 proteins were recovered in 2–7 spots; thus 51 proteins exhibiting differences between at least two treatment groups were identified overall.

View this table:
Table 3.

Uniquely identified protein spots from sham and I/R-treated ground squirrel intestines

Thirty of the protein spots that were unambiguously identified differed seasonally, i.e., between SA and IBA: 16 of these were increased in IBA, and 14 of them decreased relative to SA. Two of the proteins that increased in IBA were found in two spots, and one of the proteins that decreased was found in two spots, thus 27 distinct proteins were identified that distinguish SA from IBA intestine (Table 4). None of the proteins that differed in two spots showed the reciprocal pattern that would suggest a regulated posttranslational modification of the protein between these two states; rather, both spots were either elevated (ALB, APOA1) or decreased (leucine aminopeptidase) in IBA compared with summer. The proteins that differed during hibernation represent a broad range of functions, including facilitators of protein folding and secretion, effectors of cytoarchitecture, transport proteins, and metabolic enzymes (Table 4).

View this table:
Table 4.

Intestinal proteins that differed between SA and IBA (i.e., seasonally)

I/R treatment affected the intestines of SA animals differently than it did the intestines of IBA animals. The proteins in 18 spots that differed between SA and SA-IR intestines were unambiguously identified. Three of these proteins were found in two to six spots; therefore, 11 proteins responded to I/R in the summer intestine. As with the seasonal comparison, all three proteins changed in the same direction in all of their spots. Three of the proteins decreased upon I/R treatment in the summer animals, whereas the other eight proteins all increased (Table 5). In contrast, between IBA and IBA-IR, all seven of the protein spots that were identified decreased in the I/R-treated intestines. These seven spots included two that contained the same protein; thus six distinct proteins differentiated the I/R-treated intestines in IBA animals (Table 6). The proteins that differed between SA-IR and IBA-IR in 12 spots were identified. Five of these were also different between SA and IBA; because the fold and direction of change in the SA-IR to IBA-IR comparison mirrored the differences between SA and IBA, those were taken to have their origin as seasonal differences. Similarly, three protein spots that also differed between SA and SA-IR were taken to represent primarily the effect of IR in summer animals. Three spots remained that differed between SA-IR and IBA-IR, two that increased in SA-IR and one decreased, compared with IBA-IR (Table 7). Closer inspection reveals a difference in the patterns of these spots (Fig. 3). Spot 1729 increased in SA-IR compared with SA, but decreased in IBA-IR compared with IBA; it is only the combination of these effects that leads to a significant difference between groups. In contrast, spot 1841 increased after I/R in both SA and IBA animals, but is also lower in IBA compared with SA. In this case (and with spot 1844), the seasonal difference, together with the effect of I/R, led to a significant difference between SA-IR and IBA-IR. Variability in some sample groups (e.g., SA for spot 1841) obscures the seasonal difference and a difference between SA and SA-IR.

Fig. 3.

Abundance patterns of two representative protein spots. These spots differed significantly between summer active-after I/R treatment (SA-IR) and IBA after I/R treatment (IBA-IR) animals. Spot 1729 (A) shows opposite responses to I/R by SA and IBA ground squirrels, whereas spot 1841 (B) is increased by I/R treatment in both SA and IBA.

View this table:
Table 5.

Intestinal proteins that differed in SA animals after I/R treatment

View this table:
Table 6.

Intestinal proteins that differed in IBA animals after I/R treatment

View this table:
Table 7.

Intestinal proteins that differed between SA-IR and IBA-IR

The remaining 20 proteins identified were from spots that differed only in the cross-comparisons between sham and I/R-treated intestines, i.e., between SA-IR and IBA, or SA and IBA-IR. These differences likely reflect the combination of seasonal and I/R effects on the proteome, leading to statistically significant differences; because the biological significance of these pairs is unclear, the proteins in these spots will not be discussed further.

There are few shared protein spots among those that change primarily due to season and those that change in response to I/R treatment, and none of the protein spot changes due to I/R are shared between the SA-IR and IBA-IR treatment groups (Tables 47). Gp96 is increased in IBA and by I/R treatment in SA. ALB (spot 731), heat shock 70-kDa protein 8 (HSPA8), mercaptopyruvate sulfurtransferase, and gelsolin (GSN) are all seasonally increased in IBA but decreased by I/R treatment of IBA animals. Only one protein, ALB, appeared on both the SA and IBA lists of changes after I/R treatment; however, the two spots that were increased by I/R in the SA intestines are distinct from the spot that was decreased by I/R in the IBA group.

Three proteins identified in this proteomics screen were subjected to Western blot analysis to verify the results obtained by DIGE analysis using a more widely used method for comparing protein levels as a function of hibernation state. APOA1, UCHL1, and ALB were all expressed at higher levels in intestinal protein extracts of sham IBA vs. sham summer squirrels as expected, based on the DIGE analysis (Fig. 4). This finding validates the use of DIGE analysis as a robust, unbiased means to achieve high-throughput identification of proteomic changes associated with the hibernating phenotype and the response to I/R.

Fig. 4.

Western blot analysis of intestinal proteins in sham-treated hibernating (IBA) and SA ground squirrels. A, C, and E: representative immunoblots of apolipoprotein A-I (APOA1), albumin (ALB), and ubiquitin carboxyl-terminal hydrolase L1 (UCHL1) with bands at the indicated molecular size based on markers run on the same gel. For ALB and UCHL1, only the bands shown were detected; for APOA1, the band shown was the major band detected and the only one near the correct apparent molecular mass. B, D, and F: graphs plot mean and 1 SD obtained after quantifying protein bands in sham-treated summer and IBA animal groups. B: APOA1 (2.6-fold increased in IBA, n = 6; P = 0.001); D: ALB (2.7-fold increased in IBA, n = 5, P = 0.00046); F: UCHL1 (2.08-fold increased in IBA, n = 6, P = 0.003), Student's t-test.


The intestine of hibernating ground squirrels is significantly protected from damage in a classic model of I/R injury. This resistance is not dependent on the low Tb of torpor, because it occurs in euthermic hibernators during natural IBA. Intestinal damage by I/R is profound in summer ground squirrels compared with IBA hibernators (Fig. 1; see also Ref. 20). We predicted that changes in the proteome between summer and hibernating animals are at least partially responsible for the protected hibernation phenotype, and that significant pathways underlying this protection would be revealed by the quantitative proteomics approach of this study. We further expected that pathways involved in the damage response of the SA intestine to I/R would be revealed. The methodology used generally assessed only the most abundant, soluble proteins and thus was not optimal for revealing differences in nonabundant or integral membrane proteins. A thorough evaluation of changes in these important classes of proteins will likely be required to fully understand the molecular basis of the hibernator's protected phenotype.

This study identified a large number of protein differences between the summer and IBA sham groups; these are seasonal differences. This result is consistent with the idea that differential gene expression is an important component of hibernation (31), as well as with several reports demonstrating seasonal differences in specific mRNAs and proteins as animals transit across their circannual cycle (Ref. 2, reviewed in Refs. 8, 12, see also more recent work by Refs. 17, 42, 44). The proteins that are seasonally altered in intestine reflect various aspects of the winter phenotype, including some associated with the changes in nutrition and feeding behavior that accompany hibernation, and others that may play a role in conferring protection against the extremes encountered during the winter hibernation phase of the circannual rhythm.

3-Hydroxy-3-methylglutaryl-coenzyme A synthase 2 (HMGCS2), a key enzyme that regulates ketone body production, was expressed at higher levels in IBA vs. summer intestine. Although ketone body synthesis occurs predominantly in the adult liver, fasting induces HMGCS2 in rodent intestine (30). The increased HMGCS2 in IBA animals suggests an increased capacity for intestinal ketogenesis during hibernation, which could contribute to the known increase in circulating ketone bodies (18). Adequate energy supply to the gut is essential to support enterocyte proliferation, which increases during periodic arousal to euthermia (9).

Several proteins involved in small-molecule metabolism are decreased in IBA compared with SA. Among these are enzymes important for protein or amino acid catabolism, including peptidase D, arginase II, leucine aminopeptidase 3, and isovaleryl coenzyme A dehydrogenase. These decreases, in addition to the decrease of β-ureidopropionase, an enzyme involved in nucleotide metabolism, likely reflect the conservation of proteins and more general recycling of nitrogen that is known to occur during hibernation (see Refs. 28, 42, and references therein).

UCHL1 increased approximately twofold over summer in the IBA sham samples based on DIGE analysis and confirmed by Western blotting. This protein is strongly seasonally regulated, with significant differences seen in all four seasonal comparisons. UCHL1 is a ubiquitin hydrolase, one of a group of deubiquitylating enzymes that play crucial roles in all aspects of posttranslational regulation mediated through ubiquitylation, including protein stability, localization, and activity (24). UCHL1 is highly expressed in neurons, and, because our whole thickness intestinal samples included enteric neurons, this result may reflect relative enrichment of the enteric nervous system in hibernation. UCHL1 has been shown to have ubiquitin ligase activity in addition to ubiquitin hydrolase activity in vitro, and, although its exact substrates and function are still unknown, several studies indicate its importance in maintaining a precise balance of ubiquitin levels (1). Protein ubiquitin conjugates increase in intestine and liver during torpor (36), apparently due to a temperature-sensitive block in the proteolytic activity of the proteosome at the low Tb values of torpor (38). A second protein involved in protein ubiquitylation, the ubiquitin-activating enzyme E1, was also increased in IBA compared with SA, yet proteins are generally preserved during hibernation. It is possible that hibernators use protein mono-ubiquitylation to protect proteins during torpor (39), then rely on elevated levels of ubiquitin hydrolase, such as the UCHL1 identified here, to help reduce proteolysis by reversing the modification before the target proteins can be polyubiquitylated and degraded during rewarming.

APOA1 was elevated ∼2.7-fold in IBA vs. summer sham animals based on the DIGE analysis; this value agrees well with the 2.6-fold change obtained by Western blot analysis (Fig. 4). APOA1, the major protein component of high-density lipoprotein (HDL), is synthesized by both the intestine and the liver. Its mRNA was also found to be seasonally elevated during hibernation in golden-mantled ground squirrel liver (13). APOA1 has potent anti-atherogenic and anti-inflammatory effects; upregulation of endogenous APOA1 raises levels of HDL and inhibits the progression of cardiovascular disease in mice (see Ref. 26, for review). It is also noteworthy that administration of reconstituted HDL particles before ischemia reduced inflammatory cell infiltration and histological injury after intestinal I/R (10) in a rat model that was similar to our model in ground squirrels. Along with hepatic APOA1 synthesis, increased intestinal synthesis of APOA1 may contribute to elevated circulating levels of HDL during hibernation and help reduce the risk of atherosclerotic lesions at a time when circulating lipids are relatively abundant. There is little information on seasonal changes in lipoprotein levels in hibernators. One report found no significant difference in HDL levels between ground squirrels in the prehibernating (fattening) phase vs. hibernation, although levels earlier in the active season before fattening commences were not examined (27). Fasting increases APOA1 mRNA by 1.7-fold in young mice, although it has no effect in older animals (4). Thus, as for ALB, changes in intestinal expression of APOA1 may reflect, at least in part, the long-term fast that occurs during the hibernation season.

Two proteins involved with protein folding, the ε-subunit of chaperonin containing t-complex polypeptide 1 (CCT5), a component of the t-complex polypeptide 1 ring chaperonin complex and chaperonin (heat shock 60-kDa protein 1), as well as peroxiredoxin 5, which is an antioxidant enzyme, were expressed at lower levels in the sham IBA compared with the sham summer squirrels. Decreases in these and tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein, β-polypeptide (YWHAB), appear paradoxical in light of the protected phenotype of the hibernating intestine, but may be a general reflection of a suppressed or altered damage response during hibernation in intestine (14). Decreased expression of “protective” proteins during hibernation is not unique to the intestine, e.g., in a proteomic screen of golden-mantled ground squirrel liver, heat shock protein 60 and peroxiredoxin 3 were reduced in hibernators compared with that in summer, although other proteins with related functions were increased (12). Increased expression of HSPA8 and tumor rejection antigen (gp96, HSP90B1) in IBA compared with summer, which may be protective, were also found in this study.

There are parallels between the protein changes identified in sham-treated summer and hibernating intestines and the changes in gene expression that occur as enterocytes differentiate and migrate along the crypt-villus axis. The mRNA expression patterns for three of the proteins identified in this study as increased in IBA [APOA1, villin 1 (VIL1), and annexin A13; Table 4] were analyzed by in situ hybridization in the mouse intestine; expression of all three increased as enterocytes migrate from crypt to villus, particularly in the region containing actively absorbing cells (25). Increased UCHL1, ubiquitin-activating enzyme E1, oxoglutarate dehydrogenase, and GSN, in addition to APOA1 and VIL1, were found in to be elevated in the enriched villus population from mouse intestine, whereas expression of chaperonin containing TCP1, η (CCT7), heat shock 60-kDa protein 1, aldehyde dehydrogenase 1 family, member B1, and IVD, which decreased in IBA, were decreased (32). Villus enterocytes of hibernators tend to be hypermature in the hibernating intestine compared with those in summer animals, because proliferation and migration along the crypt-villus axis is severely depressed during torpor, increasing only for brief periods during IBAs to euthermia (9). Thus villus enterocytes can be several weeks old in hibernators, compared with the normal lifespan of 3–5 days in continuously homeothermic mammals. The available information about gene expression in the crypt-villus axis is consistent with the interpretation that hibernators have a greater proportion of mature, absorptive villus cells compared with summer animals.

The greatest fold change observed in this study was in one of several spots identified as ALB. This particular spot, 731, was elevated by nearly 15-fold in the IBA sham group compared with summer, whereas Western blot analysis revealed an ∼2.7-fold increase in intestinal ALB in sham IBA squirrels vs. summer. The Western values are not directly comparable with the values obtained in the DIGE experiment, however, because the single ALB band on the Western is an average of the multiple ALB forms that were separated into distinct spots on the 2D gels. In the 2D gel shown in Fig. 2, the ALB spot with the largest difference between summer and IBA sham samples (spot 731) is one of several ALB spots and is significantly less abundant than two of the others (i.e., the 2 spots to its immediate left); hence it is not surprising that the average fold change measured by Western blotting is significantly less than measured by DIGE analysis for spot 731 alone. ALB spot 721 was also elevated in IBA compared with SA, but it too is minor relative to several other ALB spots, again consistent with the results obtained with the Western blots. Two different ALB spots, 761 and 698, are increased by I/R in summer animals, but not in the IBA hibernators, yet I/R treatment of the IBA intestine causes a decrease of the seasonally elevated spot 731. These spots migrate as if they are all approximately the same size, but are distinguished by their charge differences. Thus these data indicate that intestinal ALB increases in winter, and that I/R causes different posttranslational modifications in summer compared with winter animals. Because ALB has been historically recognized as a protein synthesized by the liver and secreted into plasma, its increase in intestinal tissue of hibernating ground squirrels was surprising. However, extrahepatic expression of ALB mRNA has been reported recently in several tissues, including intestine (reviewed in Ref. 29). The functional significance of intestinal ALB expression is not well understood, but it may play a role in innate defense, as has been suggested for the mammary gland (29). ALB has the ability to scavenge free radicals and reduce lipid peroxidation, and thus has antioxidant and anti-inflammatory properties (7), which may contribute to the hibernation-induced protection of the mucosa after I/R (20). Elevated levels of intestinal ALB in hibernating ground squirrels may also be related to the absence of food intake during hibernation, as increased intestinal ALB has been observed during fasting in mice (21).

There were also protein differences in both SA and IBA animals following I/R, although the specific proteins that changed differed; only decreases were seen in the hibernators, whereas most of the differences in summer were due to increases in specific protein spots. The increased proteins in SA-IR could either represent a damage response that is suppressed during hibernation, or, alternatively, they may reflect the change in the relative proportions of the different intestinal cell types that were observed as a consequence of I/R. Given the substantial changes in intestinal morphology observed in summer animals after I/R treatment (Fig. 1), and the short duration of the insult (20-min ischemia followed by 60-min reperfusion), the latter mechanism is more likely to be responsible for changes in the intestinal proteome after I/R in the summer squirrels.

The majority of protein spots differentially expressed in sham vs. I/R-treated intestine in summer animals were increased after I/R (14 spots with 8 distinct proteins); only four spots decreased, with two of these containing the same protein. The I/R intestine has relatively fewer villus enterocytes than crypt cells and a greater proportion of immune cells that have infiltrated the gut in response to the I/R injury (Fig. 1, see also Ref. 20). Decreased VIL2 (ezrin) in the I/R-treated group is consistent with a relative enrichment of crypt cells in these intestines compared with those from the sham-treated intestines, as is the increase in CCT7 (32). Probably the most striking finding to emerge from the comparison of sham and I/R-treated summer animals is the significant increase in six protein spots containing transferrin that are resolved on 2D gels upon I/R treatment. Transferrin is an iron-binding plasma protein that is imported into crypt epithelial cells via the transferrin receptor under iron-replete conditions and is thought to modulate dietary iron uptake via transcriptional regulation of iron transport proteins. Increased transferrin protein in I/R-treated intestine may, therefore, reflect the relative enrichment of the I/R samples with crypt cells. Thus changes in protein expression observed for the I/R vs. sham summer comparison likely originate with the altered cell populations that characterize these different treatment groups rather than changes in the biosynthesis or stability of the proteins in static cell populations. In the response of summer intestine to I/R damage, however, the enrichment derives from a relative loss of mature villus enterocytes, in contrast to the summer to IBA differences in sham-treated animals, where the mature villus enterocytes are relatively enriched.

Strikingly, none of the protein changes in the I/R-treated IBA animals was the same as those seen in the I/R-treated summer animals. All of the proteins that changed in the IBA animals in response to I/R decreased, including four proteins shown above to be seasonally increased during hibernation (ALB, HSPA8, MPST, and GSN). Three of these are enriched in a specific subset of the villus epithelial cells (ALB, HSPA8, and GSN); thus their decrease may again reflect some loss of these cells due to the I/R treatment.

Perspectives and significance.

This proteomic screen was limited to the most abundant, soluble proteins of intestinal tissues from summer and hibernating ground squirrels; nevertheless, it identified several differentially expressed proteins that characterize the hibernating phenotype in this organ. Some of these may be particularly important for maintenance of gut integrity during the altered nutritional status of hibernating animals (e.g., HMGCS2), whereas others may play roles in whole body homeostasis (e.g., APOA1), and some may be responsible for hibernation-induced protection of the gut during experimental I/R injury (e.g., ALB). Changes in additional proteins whose abundance in the tissue is too low for detection in our proteomic screen also likely contribute to hibernation-induced protection from I/R injury or mechanistically to cell loss. These include seasonal upregulation of antiapoptotic proteins (14) and/or anti-inflammatory proteins, such as IL-10 (19). To our knowledge, this is the first report of differential protein expression in a model of intestinal I/R injury in any species. Consistent with our laboratory's previous pathophysiological study (20), unique protein changes occurred during I/R in the vulnerable summer intestine that were not shared with the protected hibernating intestine. Finally, our results highlight the importance of altered cell populations in interpreting the differences revealed by nonbiased screening methods between sham and I/R-treated animals.


This work was supported by Defense Advanced Research Projects Agency W81XWH-05-2-0016 to S. L. Martin and H. V. Carey.


We thank Dr. B. Mackenzie for helpful discussion.

Present address of C. C. Kurtz: Digestive Health Center of Excellence, University of Virginia, 409 Lane Rd., Charlottesville, VA, 22908.


  • 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