Peptidomics of Cpefat / fat Mouse Hypothalamus EFFECT OF FOOD DEPRIVATION AND EXERCISE ON PEPTIDE LEVELS *

Carboxypeptidase E is a major enzyme in the biosynthesis of numerous neuroendocrine peptides. Previously, we developed a technique for the isolation of neuropeptide-processing intermediates from mice that lack carboxypeptidase E activity (Cpe mice) due to a naturally occurring point mutation. In the present study, we used a differential labeling procedure with stable isotopic tags and mass spectrometry to quantitate the relative changes in a number of hypothalamic peptides in Cpe mice in two different paradigms that each cause an 10% decrease in body mass. One paradigm involved a 2-day fast under normal sedentary conditions (i.e. standard mouse cages); the other involved giving mice access to an exercise wheel for 4 weeks with free access to food. Approximately 50 peptides were detected in both studies, and over 80 peptides were detected in at least one of the two studies. Twenty-eight peptides were increased >50% by food deprivation, and some of these were increased by 2to 3-fold. In contrast, only three peptides were increased >50% in the group with exercise wheels, and many peptides showed a slight 15–30% decrease upon exercise. Approximately one-half of the peptides detected in both studies were identified by tandem mass spectrometry. Peptides found to be elevated by food deprivation but not exercise included a number of fragments of proenkephalin, prothyrotropin-releasing hormone, secretogranin II, chromogranin B, and pro-SAAS. Taken together, the differential regulation of these peptides in the two paradigms suggests that the regulation is not due to the lower body weight but to the manner in which the paradigms achieved this lower body weight.

Neuropeptides are involved in a wide variety of physiological processes, including reproduction, growth, anxiety, stress, sleep/wake cycles, pain, feeding and body weight regulation, and many other pathways (1).The biosynthesis of most neuropeptides requires the selective proteolysis of larger precursors, usually at sites containing multiple basic amino acids, and then removal of these basic residues by a carboxypeptidase (2).For some neuropeptides, additional steps such as acetylation, phosphorylation, sulfation, glycosylation, and/or C-terminal amidation occur; these additional modifications often alter the biological activity of the peptide (3,4).
The major peptide-processing endopeptidases are prohormone convertase (PC) 1 1 and 2 (PC1 is also known as PC3), both of which are broadly expressed in the neuroendocrine system and cleave peptide precursors to the C-terminal side of an Arg or Lys residue (5)(6)(7).This basic residue is then usually removed from the C terminus of the peptide-processing intermediate by carboxypeptidase E (CPE) (8,9).Mice lacking CPE activity due to a point mutation in the coding region of the gene (Cpe fat⁄fat mice) show reduced levels of the mature forms of neuroendocrine peptides (10 -15); this implies that CPE plays a major role but that a second carboxypeptidase also contributes to peptide processing in the absence of CPE activity.Carboxypeptidase D (CPD), which is present in the trans-Golgi network and is able to remove C-terminal Lys and Arg residues from a number of peptides, is presumably the enzyme that contributes to neuroendocrine peptide processing in the Cpe fat⁄fat mice (16 -19).However, even though CPD is present in neuroendocrine cells, the absence of CPE activity leads to a very large increase in the levels of the Lys-and Argextended intermediates in Cpe fat⁄fat mice (15).This increase in the C-terminally extended peptides led to a scheme for the purification of neuropeptide-processing intermediates from Cpe fat⁄fat mouse brain and pituitary (20), described in more detail below.
Studies on neuropeptides in brain are limited by the low levels of these peptides relative to the high levels of protein degradation fragments.Although radioimmunoassays can provide useful information about levels of neuropeptides, it is difficult to interpret the results, because antisera typically cross-react with larger or smaller forms of the peptide being measured and/or with forms containing various posttranslational modifications.We previously developed a strategy to affinity purify neuropeptide-processing intermediates from Cpe fat⁄fat mouse brain and pituitary using anhydrotrypsin columns (20).This resin specifically binds peptides containing C-terminal Lys or Arg residues, which accumulate in Cpe fat⁄fat mice.By analyzing the affinity-purified material on liquid chromatography columns coupled to an electrospray ionization mass spectrometer (LC/MS) and comparing the results obtained with Cpe fat⁄fat versus wild type mouse brain extracts, it is straightforward to determine which peptides are highly enriched in Cpe fat⁄fat mice (20).Then, tandem mass spectrometry (MS/MS) can be used to obtain sequence information and conclusively identify the peptides.Previously, these enriched peptides were found to represent either neuropeptide-processing intermediates, other fragments of proteins known to be processed in peptide-containing secretory vesicles, or novel peptides that were subsequently found to be processed in the secretory pathway (20).Recently, we have modified this basic approach to include differential isotopic labeling so that relative quantitation can be performed (15,(21)(22)(23).
In the present study, we used this quantitative peptidomics approach to examine changes in the relative levels of hypothalamic peptides in response to body weight reduction by two different paradigms.One paradigm involved food deprivation for 48 h, which resulted in a 10% reduction in body weight.The other paradigm involved voluntary exercise on activity wheels for a period of 4 weeks.At the end of this period, Cpe fat⁄fat mice given access to wheels gained less weight than sedentary Cpe fat⁄fat mice housed in standard cages, and the difference in body weight between exercising and sedentary groups was ϳ10% (24).Interestingly, even though both methods resulted in mice that were 10% lighter than control Cpe fat⁄fat mice, there were dramatic differences in the relative levels of many hypothalamic peptides in the food-deprived animals, and these changes were not observed in the exercising animals.
Animal Treatment-Several breeding pairs of BKS.HRS-Cpe fat /J mice (Cpe fat/ϩ ) were obtained from The Jackson Laboratory (Bar Harbor, ME) and maintained in the barrier facility at Albert Einstein College of Medicine.Mice homozygous for the fat mutation (i.e.Cpe fat⁄fat mice) as well as wild type littermates were obtained from crosses of the heterozygous breeding pairs.Unless indicated below, animals were housed in standard mouse cages with rodent chow (5058, Labdiet, St. Louis, MO) and water provided ad libitum and light from 7:00 to 19:00.All experimental protocols were approved by the committee for animal experimentation of the Albert Einstein College of Medicine.
For the analysis of the effect of food deprivation on hypothalamic peptides, a total of 20 Cpe fat⁄fat mice and 14 wild type littermates were used, as shown in Fig. 1.The Cpe fat⁄fat mice were divided into four groups of five animals.Each group contained either two or three females and had an average age of 15 weeks.The number of male and female mice in each group was balanced so that the two groups that were eventually pooled (Fig. 1) had an equal number of male and female mice.Two groups were deprived of food for 48 h, while the other two groups had free access to food.Wild type mice were divided into two groups of seven animals each, with three males and four females per group.One group was deprived of food for 48 h, and the other was allowed access to food.Mice were sacrificed with CO 2 vapor, and the hypothalami were dissected.Tissue was stored at Ϫ70 °C until analysis, as described below.
For the study involving voluntary exercise, all mice were 11-week-old males at the start of the 4-week study.Two groups of 3-4 Cpe fat⁄fat mice and one group of 7 wild type mice were housed individually in standard mouse cages (28 ϫ 18 cm floor, 12-cm height) without access to exercise wheels ("sedentary" group).Another two groups of 3-4 Cpe fat⁄fat mice and one group of 7 wild type mice were housed individually in standard rat cages (48 ϫ 27-cm floor, 20-cm height) equipped with exercise wheels (14.5-cm diameter).Both sedentary and exercising groups were allowed free access to food and water.Body weight was measured every 2 days, as described (24).After 4 weeks, the animals were sacrificed by CO 2 asphyxiation, and hypothalami were dissected and stored at Ϫ70 °C until analysis.
For the analysis of the effect of food deprivation on mRNA levels, a total of 16 mice were used: 8 Cpe fat⁄fat mice and 8 wild type littermates, all 13-15 weeks of age.Four animals of each genetic background were deprived of food for 48 h, while the other four animals of the group were allowed free access to chow.Mice were then sacrificed, and hypothalami were removed.RNA was prepared using the RNeasy Mini kit (Qiagen), quantified by absorbance at 260 nm, and stored at Ϫ80 °C.Equal amounts of RNA from each animal were analyzed separately on Northern blots, as described below.
Peptide Extraction and Labeling with Differential Stable Isotopic Tags-To extract peptides from Cpe fat⁄fat or wild type mouse hypothalami, 300 l of 10 mM HCl was added for each hypothalamus in the tube (Fig. 1).The tissue was sonicated three times each for 5 s and incubated at 70 °C for 20 min.The homogenate was combined with 200 l of 0.2 M phosphate buffer, pH 9.5, and centrifuged at 50,000 ϫ g for 40 min at 4 °C.The supernatant was taken out, and the pellet was resuspended in 100 l of 0.2 M phosphate buffer, pH 9.5, and centrifuged.The second supernatant was combined with the first and concentrated to 500 l in a vacuum centrifuge.Another 200 l of 0.2 M phosphate buffer, pH 9.5, was added to the supernatant, and the pH was adjusted to 9.5 with 1 M NaOH.To label the resulting mixture, 6 l of 300 g/l TMAB reagent in Me 2 SO was added to the mixture.The extract from one group of fed or sedentary animals was labeled with the H 9 -TMAB reagent, and the extract from a food-deprived or exercising group was labeled with the D 9 -TMAB reagent, as indicated (Fig. 1).Another two groups were labeled with the reverse scheme, so that the food-deprived or exercising group received the H 9 -TMAB reagent while the control group received the D 9 label.After 10 min, the pH was adjusted back to 9.5 with 1.0 M NaOH.The addition of TMAB and NaOH was repeated six times over 4 h.After adjustment of the pH to 9.5 again, 80 l of 2.5 M glycine was added to the reaction mixture to quench any remaining label reagents.After 40 min with glycine, the light and heavy reagent-labeled samples were pooled as shown in Fig. 1 and filtered through a Microcon YM-10 unit (Millipore, Bedford, MA) to remove molecules with a molecular mass greater than 10 kDa.After the pH was adjusted to 9.0 -9.5, 2.0 M hydroxylamine (in Me 2 SO) was added to remove any labels that might have attached to the hydroxyl side chains of tyrosine, serine, or threonine residues.After repeating this step twice, 1.0 M NaAc, pH 5.0, 5% CHAPS, and 1.0 M CaCl 2 were combined with the reaction to obtain a final concentration of 50 mM, 0.25%, and 20 mM, respectively.The pH was adjusted to 5.0 with 6 M acetic acid.The resulting mixture was subjected to affinity purification on a column containing ϳ0.5 ml anhydrotrypsin-agarose (60 -90 nmol of protein/ml of resin) to selectively isolate peptide containing C-terminal basic residues.The column was first washed with 28 ml of 0.25% CHAPS in 50 mM NaAc buffer, pH 5.0, containing 20 mM CaCl 2 and 0.5 M NaCl and then rinsed with 32 ml of 10 mM NaAc buffer, pH 5.0, containing only 10 mM CaCl 2 .Peptides were eluted first with 6 ml of water, then with 12 ml of 6 mM HCl.The eluates were combined and concentrated to 100 l in a vacuum centrifuge.Aliquots (50%) of sample were used for further analysis, as described below.
Liquid Chromatography and Tandem Mass Spectrometry-LC/ MS/MS experiments were performed on an Applied Biosystems/MDS SCIEX API QSTAR PULSARi TM quadrupole time of flight mass spectrometer coupled with a micro-ESI source.An UltiMate TM Capillary/ Nano LC System was directly connected to the mass spectrometer.After loading on a PepMap TM C 18 trapping column (5 m, 100 Å, 300 m i.d.ϫ 5 mm, LC Packings) and desalting for 20 min with 2% acetonitrile in 0.1% formic acid, the sample was backflushed into and separated on a Vydac MS C 18 capillary column (5 m, 300 Å, 75 m i.d.ϫ 15 cm) at a flow rate of 200 nl/min with a gradient from 2% to 52% solvent B in 45 min, then to 90% B in 20 min.Solvent A was 2% acetonitrile and 0.1% formic acid in water.Solvent B was 80% acetonitrile and 0.1% formic acid in water.LC/MS/MS was carried out with the Information Dependent Acquisition method.The ion spray voltage of the micro-ESI source was 5.5 kV.MS/MS spectra were acquired for 2 s, and the dynamic exclusion time for previously fragmented ions was 300 s.The Rolling Collision Energy feature of Analyst QS software was used for the collision-induced dissociation (CID) of peptide ions.
All m/z ions detected in the LC/MS analysis of the Cpe fat⁄fat mice were used to search for ions with the same mass, charge, and elute time in the wild type extracts.Further analyses (i.e.quantitation and identification) were further performed only for these peptides found in the Cpe fat⁄fat extracts and not present in the wild type extracts; previous studies have shown that those peptides unique to Cpe fat⁄fat mice represent neuroendocrine peptides and other peptides generated in the secretory pathway (20,25).Relative levels of peptides in treated versus non-treated Cpe fat⁄fat mouse hypothalami were determined by the ratio of peak intensity of the H 9 -TMAB-and D 9 -TMAB-labeled peptide pairs, as described (15,23).All MS/MS spectra of labeled peptides were interpreted manually.The experimentally determined parent mass of the TMAB-labeled peptide was converted to the parent mass of the unlabeled and unprotonated peptide using the isotopic masses of the light (128.1177)and heavy (137.1728)forms of TMAB label as described previously (23).For a peptide to be considered as "identified," the parent mass of a peptide had to be within 30 ppm (preferably 20 ppm) of the theoretical value and more than 80% of the CID fragments observed in the MS/MS spectra had to match predicted fragments, with at least five matches to b-and/or y-ions.
Northern Blot Analysis-RNA samples were fractionated on 1.2% w/v agarose gels containing formaldehyde.RNA was transferred to a Hybond-N nylon membrane (Amersham Biosciences) by capillary action with 10ϫ SSC and fixed by baking for 2 h at 80 °C in a vacuum oven.To generate a SAAS-probe for Northern blotting analysis, XbaI-linearized pBluescript KS vector (Stratagene) carrying an 835-bp PCR fragment of mouse pro-SAAS cDNA, T7 RNA polymerase, and [␣-32 P]UTP were used.For the proenkephalin probe, a p64E vector carrying a 1070-bp fragment of mouse proenkephalin cDNA was cut with EcoRI, and the linearized vector was used as a riboprobe template with SP6 RNA polymerase and [␣-32 P]UTP.To create the CPD probe, pGEM72 vector carrying a 600-bp mouse CPD PCR product (corresponding to carboxypeptidase domain 2) was cut with BamHI and the linearized product used in the riboprobe reaction with T7 RNA polymerase and [␣-32 P]UTP.The blots were prehybridized in hybridization buffer (50% v/v deionized formamide, 6ϫ SSC, 0.5% SDS, and 100 g/ml salmon sperm DNA) for 1 h at 60 °C prior to addition of [␣-32 P]UTP-labeled probes (12-15 ϫ 10 6 cpm per 20 ml of hybridization buffer).After hybridization overnight at 60 °C, the blots were washed twice with 2ϫ SSC containing 0.1% SDS at 60 °C for 30 min and twice with 0.2ϫ SSC containing 0.5% SDS at 60 °C for 30 min.Labeled bands were detected by autoradiography at Ϫ70 °C with Kodak film.Quantitation was performed by densitization of the autoradiogram and normalization to the level of either 18 S or 28 S rRNA.

RESULTS
Altogether, 82 different peptides were detected upon LC/MS analyses of the extracts of Cpe fat⁄fat mouse hypothalamus that were not detectable in the extracts of wild type mouse hypothalamus (Table I).Previously, those peptides found to be greatly enriched in Cpe fat⁄fat mouse brain extracts, relative to wild type mouse brain extracts, were neuropeptide-processing intermediates or other fragments of neuroendocrine secretory pathway proteins (20).Of the 82 peptides detected in the present study, 40 were identified by MS/MS sequencing (Table I).Representative MS/MS data are shown in Fig. 2. The criteria for considering a peptide identified include a match of the parent mass within 30 ppm of the predicted mass, and at least five fragment ions that matched a predicted b-or y-series ion, also within 30 ppm.An additional nine peptides matched the theoretical parent mass of a known hypothalamic peptide, but there was insufficient MS/MS data to permit a conclusive identification; these nine peptides are indicated in Table I with parentheses.Thirty-three other peptides that were detected upon LC/MS of the Cpe fat⁄fat mouse hypothalamic extracts could not be assigned, either because the ions gave very low signals that were not selected by the mass spectrometer for MS/MS sequencing or because the MS/MS spectra could not be interpreted.Uninterpretable spectra could be due to a low signal strength, which results in few MS/MS fragments, a large peptide size that produces a complex MS/MS fragmentation pattern, and/or post-translational modifications that alter the parent and fragment masses.All of the 40 peptides identified from MS/MS sequencing as well as the 9 additional peptides that were tentatively assigned based only on parent mass were modified with the appropriate number of isotopic tags: one tag per free N terminus and one per Lys residue (Table I).This provides an additional confirmation that the identification of the peptide is correct.Many of the 40 peptides identified in the present study match those found previously using a similar affinity column-based approach, but without the TMAB isotopic tagging (20).
The incorporation of the stable isotopic tags into the peptides provides a method to examine relative levels in two different groups of mice.In the present study, this technique was used to examine hypothalamic peptides in two different paradigms that affect body weight: food deprivation and exercise.When Cpe fat⁄fat mice (average age ϳ15 weeks) were deprived of food for 48 h, their body weight decreased from an average of 37.6 to 33.9 g, a loss of 10.9%.In contrast, Cpe fat⁄fat mice with free access to food increased from 39.4 to 40.1 g over the same 48-h period, an increase of 2%.For the other paradigm, one group of Cpe fat⁄fat mice was given exercise wheels starting at 11 weeks of age and then allowed free access to food for 4 weeks.A control group of Cpe fat⁄fat mice was housed in standard mouse cages for the same time, also with full access to food.This sedentary group of mice gained an average of 23% body weight during this FIG. 1. Scheme showing number of animals per group, labeling reagent, and pooling strategy for the various groups in the two experimental paradigms.Top: mice were housed in standard mouse cages and were either provided free access to food or were deprived of food for 48 h before sacrifice.Bottom: mice were given free access to food and were housed in either standard mouse cages (Sedentary) or in larger rat cages equipped with exercise wheels for 4 weeks.In both studies, the mice were ϳ15 weeks old at the time of sacrifice.Hypothalami were extracted and labeled with either H 9 -TMAB reagent (Htag) or the D 9 -TMAB reagent (D-tag).The extracts were pooled as indicated and then purified on anhydrotrypsin-agarose columns and analyzed by LC/MS/MS as described under "Experimental Procedures."

TABLE I Relative levels of hypothalamic peptides after food deprivation or exercise
The indicated peptide name (column 2) represents the name of the peptide lacking the C-terminal basic residues.Sequences (column 3) were identified from MS/MS and fit the criteria described under "Experimental Procedures," unless enclosed by parentheses; those in parentheses represent predicted peptides based on match to parent mass and, in some cases, partial MS/MS data that was not sufficient for conclusive identification.Theor.mass (column 4) indicates the theoretical monoisotopic mass of the unprotonated and unlabeled peptide, and Obs.mass (column 5) indicates the observed monoisotopic mass, after subtracting the mass of the protons and TMAB groups.The difference between the observed and theoretical masses in parts per million (ppm) is shown (column 6).The number of TMAB labels incorporated into the peptide (#T) is indicated (columns 7 and 9).Ratio E/S (column 8) is the ratio of the relative peak intensity of the signals for the mice with exercise wheels versus the sedentary mice.Results are shown for two separate analyses, and the range of these two experiments is indicated.Ratio N/F (column 10) is the ratio of the relative peak intensity of the signals for the mice that were food-deprived versus the fed mice.Results are shown for two separate analyses, and the range of these two experiments is indicated.period, with an average body weight of ϳ40.0 g at the end of the study.In contrast, the mice with exercise wheels increased from 32.7 g at 11 weeks of age to only 36 g after 4 weeks with the wheel; a gain of only 10%.Thus, the mice with exercise wheels had body weights that were ϳ10 -15% smaller than a comparable group of sedentary mice.The two paradigms therefore resulted in roughly the same difference in body weights between the control sedentary groups with full access to chow.When mice were deprived of food for 48 h, a large number of hypothalamic peptides were found to increase (Fig. 3 and Table I).To reduce the chance of artifacts that could occur if the two isotopic reagents had differential reactivities, and to reduce the error due to animal to animal variation, four groups of five Cpe fat⁄fat mice per group were used in this experiment so that two LC/MS runs could be performed; one with the fed group labeled with the H-reagent and the food deprived group with the D-reagent, the other with the fed group labeled with the D-reagent and the food-deprived group with the H-reagent (Fig. 1).In general, the two analyses for the same animal treatment showed excellent agreement (Fig. 3 and Table I).For the majority of peptides, the duplicate analyses were within 20% of each other (ranges shown in Table I, columns 8 and 10).Also, in those cases where multiple peptides were detected that arise from the same precursor through selective processing (chromogranin B, proenkephalin, pro-SAAS, pro-TRH, and secretogranin II) the majority of the peptides in each group showed generally similar changes (Table I).However, there were several exceptions to this tendency, which could be explained by differential processing of the precursors into distinct products that are regulated differently by food deprivation (discussed below).
In contrast to the results with food deprivation, most hypothalamic peptides detected in the present study did not change dramatically upon exposure of the mice to exercise wheels (Fig. 4 and Table I).The average ratio for all peptides detected in the exercise study was 0.997, whereas the similar ratio for the food deprivation study was 1.605.Although the average level of all peptides was not changed in the exercising mice, relative to sedentary mice, there were a few peptides that showed modest changes.For example, many of the proenkephalin-derived peptides showed a slight 15-30% decrease in the exercising group; many of these same peptides increased substantially upon food deprivation (Table I).Only four peptides in the exercising mice showed large changes (Ͼ50%); these include a peptide tentatively identified as vasopressin (59% decrease), another peptide identified as a C-terminal fragment of provasopressin (80% increase), and two unidentified peptides that increased 141-212%.
To determine if the mRNA for selected prohormones was altered by food deprivation in the Cpe fat⁄fat mice, Northern blot analysis was used to compare levels of various mRNAs in fasted and fed mice.In addition, similarly treated wild type mice were also examined by Northern blot analysis.The prohormone mRNAs chosen for further analysis included proenkephalin and pro-SAAS, both of which encode a large number of peptides found to be regulated by food deprivation (Table I).Despite the finding that many of the proenkephalin and pro-SAAS-derived peptides were elevated by food deprivation, neither mRNA was significantly increased by the treatment (Fig. 5).A similar result was found in both Cpe fat⁄fat and wild type mice (Fig. 5).As a side point, both proenkephalin and pro-SAAS mRNAs were expressed in Cpe fat⁄fat mice at levels similar to those in wild type mice, indicating that the extreme change in body weight caused by the mutation of CPE does not affect the hypothalamic levels of these mRNAs.Because the technique used in the present study detected peptide-processing intermediates containing C-terminal basic residues, and not the mature forms of the neuropeptides, it was important to investigate whether CPD mRNA levels were altered by the food deprivation; if so, then this could have a large effect on the levels of the Lys-and Arg-extended processing intermediates.Fasting did not significantly alter the level of hypothalamic CPD mRNA either in wild type or in Cpe fat⁄fat mice (Fig. 5).Furthermore, the level of CPD mRNA is similar in wild type and Cpe fat⁄fat mice, indicating that the CPE mutation does not lead to an increase in CPD (Fig. 5).

DISCUSSION
Most previous studies investigating the effect of food deprivation on neuropeptides have either focused on individual neuropeptides using techniques such as radioimmunoassays, or have looked at mRNA by Northern blotting or in situ hybridization.Recently, microarray analysis was used to screen for hypothalamic mRNAs that were up-or down-regulated by a 48-h fast (26).However, only 1 of the 169 genes found to be significantly up-or down-regulated by the fast encoded a neuropeptide (neuropeptide Y).The peptidomics approach used in the present study enables the analysis of a large number of peptides, many of which are the immediate precursor of known biologically active peptides.Unlike radioimmunoassays, the quantitative peptidomics method doesn't require prior knowledge about a peptide and provides information on the precise form of the peptide being measured.Although radioimmunoassays are sensitive, the antisera used are rarely specific for a particular form of the peptide and so it is usually not possible to make conclusions about the exact peptide being measured (i.e.including post-translational modifications).The use of stable isotopic labels enables the accurate determination of the relative levels of a peptide in two different samples and can analyze a large number of peptides in a single run (27,28).Previously, the TMAB isotopic labeling approach was evaluated in terms of reproducibility by splitting an extract of pituitaries into two parts, labeling each part with H 9 -or D 9 -TMAB labels, and then pooling and analyzing by LC/MS (15,23).In this analysis, all of the peptides showed a D/H ratio within 10% of the expected 1.00, and most were within 5% of this ratio.The variability between animal groups can be larger than this, due either to animal-to-animal variations or to differences in the handling of the animals or the dissections and extractions.In the present study, the pooling of groups of 3-4 animals in the exercise paradigm and 5 animals in the food deprivation paradigm was designed to reduce these variations while still allowing for two independent analyses for each paradigm.In most cases, the observed peptides showed excellent agreement between the two analyses for the same treatment (discussed further below).
One drawback to the current method is that, for analysis of peptides in brain, the technique is limited to studies on Cpe fat⁄fat mice so that the Lys-and Arg-extended peptides can be purified on anhydrotrypsin columns.Without this purification step, the levels of neuropeptides in brain extracts are so much lower than the degradation fragments of proteins that LC/MS/MS analysis of brain extracts does not detect more than a handful of neuropeptides.While some protein degradation fragments contain C-terminal Lys or Arg residues and are also purified on the anhydrotrypsin column, a comparison of the LC/MS/MS runs from wild type and Cpe fat⁄fat mice shows these nonspecific fragments to be present in both genotypes, whereas the neuropeptide-processing intermediates are detected only in the extracts from Cpe fat⁄fat mice (20).Another limitation of the technique is that MS does not detect all peptides present in an extract.Some peptides form ions more readily than others, and the relative signal strength on MS is quite variable among different peptides.For example, neuropeptide Y is considered to be one of the most abundant peptides in brain, especially hypothalamus, but MS analyses of brain conducted by our D, CID spectrum of the mono-protonated tetra-charged precursor ion with a monoisotopic m/z of 811.65.The parent mass (3245.59)matches with the predicted monoisotopic mass (3245.61) of tri-H 9 -TMAB-labeled phosphorylated proenkephalin fragment 239 -262 (FAESLP-phosphoS-DEEG-ENYSKEVPEIEKR).Note that during CID there is frequently the neutral loss of H 9 -trimethylamine (mass ϭ 59) from the H 9 -TMAB-labeled peptide and D 9 -trimethylamine (mass ϭ 68) from the D 9 -TMAB-labeled peptide, as described previously (15,23).For all four CID spectra, b n Ј ϭ b n Ϫ 59, y n Ј ϭ y n Ϫ 59, y n Љ ϭ y n Ϫ (59ϫ2), yٞ ϭ y n Ϫ (59 ϫ 3), a n Ј ϭ a n Ϫ 59, MЉ ϭ M Ϫ (59 ϫ 2), and Mٞ ϭ M Ϫ (59 ϫ 3).After neutral loss of trimethylamine, the remaining portion of the TMAB group is positively charged, so the bЈ and aЈ ions are non-protonated singly charged ions.group (15,20,21,23,25) or by others (29) have not detected this peptide.Still, it is possible to detect a relatively large number of peptides with MS, and while not inclusive of every brain peptide, the MS-based methods provide a survey of relative changes in the levels of many peptides.Some of the changes in peptides detected in the present study are consistent with previous studies that examined peptide levels in hypothalamus and/or subregions of the hypothalamus in response to food deprivation.For example, ten pro-SAAS-derived peptides were detected in the present FIG. 3. Mass spectra of TMAB-labeled peptide pairs obtained from LC/MS analysis of the fasted and nonfasted Cpe fat⁄fat hypothalamus.A and B, the tetra-charged ion with a monoisotopic peak at m/z 563.84 or 566.09 represents mono-H 9 -or D 9 -TMAB-labeled little SAAS-RR, respectively.C and D, the tetra-charged ion with a monoisotopic peak at m/z 714.67 or 719.19 represents di-H 9or D 9 -TMAB-labeled PEN-KR, respectively.E and F, the tetra-charged ion with a monoisotopic peak at m/z 513.80 or 520.59 represents tri-H 9 -or D 9 -TMAB-labeled proenkephalin fragment 198 -211.G and H, the tetra-charged ion with a monoisotopic peak at m/z 811.65 or 818.44 represents tri-H 9 -or D 9 -TMAB-labeled phosphorylated proenkephalin fragment 239 -262.For each set, the spectra on the left (A, C, E, and G) represent labeling of the fed group with the light label and the fasted group with the heavy label.The spectra on the right (B, D, F, and H) represent the reverse labeling scheme, with the fed group labeled with the heavy reagent and the fasted group with the light reagent.The ion m/z difference between the light and heavy forms of each peptide pair (⌬m) is expressed by the formula ⌬m ϭ 9.055 ϫ T/z, where T is the number of TMAB tags on the peptide and z is the total charge of the peptide (i.e. the number of protons plus the number of TMAB tags).For the ion pairs shown in this figure, z ϭ 4, so the ⌬m is ϳ2.264Da per TMAB group (1 in panels A and B, 2 in panels C and D, and 3 in the other panels).
study.Of these peptides, three showed no major change (Ͻ15%), three others showed a modest increase of 33-44%, and four peptides showed increases of 100% or more (Table I).Recently, it was reported that pro-SAAS protein levels were elevated by a 65-h fast in rat median eminence but not in paraventricular nucleus (30).Pro-SAAS-derived peptide levels were not examined in this previous study, and it is known that pro-SAAS in rat brain is largely processed into FIG.4. Mass spectra of TMAB-labeled peptide pairs obtained from LC/MS analysis for the exercising and sedentary Cpe fat⁄fat hypothalamus.Peak assignments were as in Fig. 3.For each set, the spectra on the left (A, C, E, and G) represent labeling of the sedentary group with the light label and the exercise wheel group with the heavy label.The spectra on the right (B, D, F, and H) represent the reverse labeling scheme, with the sedentary group labeled with the heavy reagent and the exercise wheel group with the light reagent.The ion m/z difference between the light and heavy forms of the peptide pairs are defined in the Fig. 3 legend.
smaller peptides with some tissue-and/or region-specific processing differences (31)(32)(33)(34).Thus, it is possible that pro-SAAS is regulated differently in subregions of the hypothalamus and that due to processing differences in these subregions the levels of individual pro-SAAS-derived peptides are regulated to a different extent by the food deprivation.This possibility is supported by the finding that the levels of pro-SAAS mRNA in hypothalamus are not significantly altered by food deprivation, either in Cpe fat⁄fat or wild type mice (Fig. 5).ProSAAS and/or pro-SAAS-derived peptides have a putative role in body weight regulation based on the observation that mice overexpressing pro-SAAS are overweight (35).The changes in pro-SAAS-derived peptides detected in the present study are consistent with the proposed role of these peptides in regulating body weight.
Many peptides derived from proenkephalin were also found to be up-regulated upon food deprivation, but without a corresponding change in the precursor mRNA levels.All eight proenkephalin-derived peptides that were detected in the food deprivation study were elevated by the treatment, with an average increase of 78% (Table I).However, hypothalamic proenkephalin mRNA levels were not significantly altered by food deprivation (Fig. 5).These observations fit with previous reports on enkephalin peptide and proenkephalin mRNA levels.Levels of immunoreactive Met-enkephalin in the ventromedial hypothalamus and the paraventricular nucleus were elevated as much as 30 -56% by food deprivation, depending on the circadian state of the rats (36), but proenkephalin mRNA in the arcuate nucleus was not affected by food deprivation of 24 or 48 h (37).
Other changes in hypothalamic peptides observed in the present study upon food deprivation are consistent with previous studies investigating neuropeptide mRNA levels.For example, in the present study five secretogranin-derived peptides were detected, all of which increased upon food deprivation (three were moderate increases of 42-44%; the other two were ϳ200% increases).Previously, secretogranin mRNA levels were increased 47% in rat paraventricular nucleus after 4 days of food deprivation (38).A single peptide derived from promelanin-concentrating hormone was detected in the present study and found to increase 47% by food deprivation (Table I).Previously, hypothalamic promelanin-concentrating hormone mRNA was reported to show a moderate increase upon food deprivation (39,40).Consistent with these observations is the finding that mice lacking promelanin-concentrating hormone are lean (41).
Although most of the observed changes in peptides detected in the present study are consistent with previous studies, peptides derived from prothyrotropin-releasing hormone (pro-TRH) showed a change opposite to that previously reported in the literature.In the present study, seven out of the nine pro-TRH-derived peptides that were detected showed an increase of 42-96%; the other two peptides showed negligible changes.However, previous studies found that pro-TRH mRNA and pro-TRH-derived peptides were decreased in rat hypothalamus upon food deprivation (30,42,43).It is not clear as to why the pro-TRH-derived peptides detected in the present study showed an increase upon food deprivation, whereas in previous studies the mature form of TRH showed a decrease by a similar treatment.It is possible that the differential regulation is due to the different species of the animals used in the various studies; rats in the previous studies on TRH and mice in the present study.Alternatively, the C-terminally extended peptides detected in the present study may be regulated independently of the mature forms for some peptides.In support of this latter possibility is the recent observation that levels of PC1 (but not PC2) protein in the paraventricular nucleus and median eminence are decreased by a 65-h fast (30).The decrease in PC1 levels would be expected to delay the rate of pro-TRH processing within the secretory pathway.Typically, peptide processing is thought to begin in the trans-Golgi net- work or immature secretory vesicles and to continue as the vesicles mature.Once PC1 or PC2 cleaves the prohormone, then CPE (or CPD) is thought to rapidly remove the basic residues from the C terminus.Because CPE is inactive in the Cpe fat⁄fat mice, any peptides generated in the late secretory pathway (which contains CPE but not CPD) would not have the basic residues removed from the C terminus.Thus, a delay in the rate of endopeptidase processing in the Cpe fat⁄fat mice may result in an accumulation of Lys-and/or Arg-extended peptides.In the present study, the absence of any detectable change in CPD mRNA levels upon food deprivation implies that alterations in CPD activity do not occur; if CPD activity was regulated, this would presumably cause changes in levels of the C-terminally extended peptides.Thus, the observed changes in levels of the Lys-and/or Arg-extended peptides detected in the present study are likely to represent real changes in the biosynthetic rate of the peptides.
Although both food deprivation for 2 days and voluntary exercise for 4 weeks resulted in Cpe fat⁄fat mice that had body weights about 10 -15% lower than Cpe fat⁄fat mice housed under sedentary conditions with free access to food, the two paradigms had much different effects on the peptides.Whereas food deprivation caused large changes in many peptides, the voluntary exercise caused a change of more than 20 -30% in only a small number of peptides.Also, the food deprivation produced an increase in the majority of peptides, whereas exercise tended to produce a slight decrease of 15-30% (Table I).There were, however, some changes noted in the exercising group.One peptide tentatively identified as vasopressin with a Cterminal Lys-Arg-extension decreased ϳ60% and a fragment of provasopressin located near the C terminus of this prohormone increased by 80% in the exercising mice.This difference presumably reflects the different processing enzymes that generate these two peptides.Vasopressin-Gly-Lys-Arg is the immediate product of PC1 or PC2 activity on provasopressin, whereas the C-terminal peptide is formed by cleavage of provasopressin at sites that do not fit the consensus for PC-like cleavages; the N terminus of the peptide requires cleavage between Leu and Val and the C terminus requires cleavage between Arg and Val (but with no upstream basic residue in the appropriate position for a PC-consensus site).Therefore, differential regulation of the various enzymes could account for the decrease in one provasopressin-derived peptide and the increase in another.In addition to these changes, two unidentified peptides were found to increase substantially upon exercise, one with a mass of 1986.13 increased 141% and another with a mass of 3303.57increased by over 200%.Unfortunately, no MS/MS sequence information was available for either of these peptides, and their identity could not be determined.Some of the peptides detected in the present study may represent previously undescribed fragments of prohormones.Although the majority of peptides result from cleavage at PCconsensus sites at both the N-and C-terminal side of the peptide, some of the observed peptides result from non-traditional processing sites.A number of these non-traditional sites represent cleavage at a hydrophobic N-terminal site.Examples include the pro-SAAS fragment of mass 1707.89(cleavage at L-G), the pro-TRH fragments of mass 1046.55 and 1694.84 (cleavages at W-F and F-F, respectively), the provasopressin fragment of mass 1712.92mentioned in the preceding paragraph (cleavage at L-V and R-V), and the secretogranin fragments of mass 1324.75, 1537.82, and 1554.85 (cleavages at Y-P, F-Q, and F-Q).Previous studies have detected cleavage at hydrophobic sites (20, 44 -46), although it is not clear what enzyme performs these cleavages in vivo.The use of mass spectrometry provides accurate information regarding the precise molecular form of the peptide being investigated, which is generally lacking in studies using radioimmunoassays to measure peptide levels.It is likely that, with increasing use of mass spectrometry to analyze neuroendocrine peptides, a larger number of non-traditional processing sites will be detected.

FIG. 5 .
FIG. 5. Northern blot analysis of hypothalamic RNA.Total RNA from mouse hypothalamus was fractionated on a formaldehyde agarose gel, transferred to a nylon membrane, and probed with 32 P-labeled probes for proenkephalin (top panel), pro-SAAS (middle panel), or CPD (bottom panel).Within each panel, the inset shows a representative autoradiogram from fed and fasted wild type mice (WT) and Cpe fat⁄fat mice (fat/fat).The signal was quantified and adjusted for the amount of 18 S or 28 S rRNA in each lane.The bar graphs show the average mRNA level; error bars indicate standard error of the mean (n ϭ 4).