Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Recently, two novel hypothalamic peptides were isolated and subsequently named orexin-A and orexin-B (Sakurai et al., 1998) or hypocretin-1 and hypocretin-2 (de Lecea et al., 1998). Despite some initial confusion, orexin-A should now be considered identical to hypocretin-1 and orexin-B to hypocretin-2. Orexins act as agonists on two G-protein-coupled receptors called OX₁ and OX₂ receptors. Increased wakefulness and reduced sleep is a well demonstrated response to central administration of orexin, and disruption of central orexinergic signaling leads to the sleep disorder narcolepsy in animal models and probably also in man (reviewed in Beuckmann and Yanagisawa, 2002; Kukkonen et al., 2002; Sutcliffe and de Lecea, 2002). The other physiological roles for orexins may be regulation of energy homeostasis and stress response, probably both via central and peripheral mechanisms (reviewed in Willie et al., 2001; Beuckmann and Yanagisawa, 2002; Kirchgessner, 2002; Kukkonen et al., 2002; Smart and Jerman, 2002).

The two orexin peptides, orexin-A and -B, are both products of the same precursor peptide, preproorexin, cleavage of which results in equimolar amounts of orexin-A and orexin-B. Orexin-A is composed of 33 amino acids and it contains two disulfide bridges, whereas orexin-B is a linear peptide of 28 residues (Sakurai et al., 1998). Although a product of a different part of the precursor peptide, orexin-B shows a 46% sequence identity with orexin-A, and these two peptide sequences seem to have arisen through duplication of a single sequence (Alvarez and Sutcliffe, 2002). Most striking is the homology in the more C-terminal parts of the peptide, which could make orexin-B an N-terminally truncated variant of orexin-A. The secondary structure of orexin-A is not known, but orexin-B has been determined to consist of two -helices in 60- to 80° angles to each other (Lee et al., 1999). Orexin-A is much more lipophilic than orexin-B, and it is also more stable in blood and cerebrospinal fluid (Kastin and Akerstrom, 1999). Yet, the CNS (central nervous system) orexin-B levels are consistently 2 to 5 times higher than orexin-A levels (Mondal et al., 1999a,b; Date et al., 2000a,b).

Orexins act as agonists on two G-protein-coupled receptors called OX₁ and OX₂ receptors. Both of these subtypes show a high (91-98%) interspecies conservation between different mammalians. Human variants of OX₁ and OX₂ receptors share a 64% sequence identity. Many studies suggest Ca²⁺ influx as the most immediate cellular response to orexin receptor activation in different systems (van den Pol et al., 1998, 1999; Lund et al., 2000; Hirota et al., 2001; Kukkonen and Åkerman, 2001; Uramura et al., 2001; Holmqvist et al., 2002). This Ca²⁺ influx may in some systems occur via protein kinase C-dependent activation of voltage-gated Ca²⁺ channels (van den Pol et al., 1998; Uramura et al., 2001; Xu et al., 2002) but a different type of Ca²⁺ channel has been implicated in other systems (Lund et al., 2000; Kukkonen and Åkerman, 2001). In the CNS, the most prominent response to orexin application is an increase in synaptic activity (reviewed in Beuckmann and Yanagisawa, 2002; Kukkonen et al., 2002). The putative role of increased Ca²⁺ influx for this process is largely unclear (reviewed in Kukkonen et al., 2002). Orexin receptor subtypes are somewhat differentially distributed (reviewed in Willie et al., 2001; Kukkonen et al., 2002; Smart and Jerman, 2002), but since there is yet no evidence of subtype selective signaling, it is difficult to predict the significance of this difference.

When Ca²⁺ or inositol phosphate responses are measured in CHO (Chinese hamster ovary), PC12, or Neuro-2a cells recombinantly expressing orexin receptors, orexin-A is 10 to 100 times more potent than orexin-B on the OX₁ receptors, whereas both orexins are equipotent on the OX₂ receptors (Sakurai et al., 1998; Smart et al., 1999; Okumura et al., 2001; Holmqvist et al., 2002). These potencies appear to be direct reflections of the binding affinities (Sakurai et al., 1998), suggesting that despite the similarities between the orexin receptors there are interesting differences in the ligand binding domains. To further examine the parameters required for the orexin peptide-receptor interaction, we have in this study investigated the determinants of orexin-A required to activate OX₁ and OX₂ receptors. Truncated and alanine-substituted orexin-A peptides have been tested for their ability to induce Ca²⁺ elevations in CHO-K1 cells heterologously expressing OX₁ or OX₂ receptors. The results suggest that the structures responding to orexin-A are highly conserved between OX₁ and OX₂ receptors, but also that there are some clear differences between these receptors.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Culture. CHO-hOX₁-C1 cells were produced as described in Lund et al. (2000). CHO-hOX₂-C1 cells were produced in a similar way; the clones used expressed receptors at approximately the same level, as determined with [¹²⁵I]orexin-A binding. CHO-hOX₁-C1 and CHO-hOX₂-C1 cells were grown in Ham's F-12 medium (Invitrogen, Carlsbad, CA) supplemented with 100 U/ml penicillin G (Sigma-Aldrich, St. Louis, MO), 80 U/ml streptomycin (Sigma-Aldrich), 400 µg/ml geneticin (G418; Invitrogen) and 10% (v/v) fetal calf serum (Invitrogen) at 37°C in 5% CO₂ in an air-ventilated humidified incubator in 260-ml plastic culture flasks (75 cm² bottom area; NUNC A/S, Roskilde, Denmark). For fluorometry, the cells were grown on circular plastic culture dishes (i.d., 94 mm; NUNC A/S).

Drugs. EGTA (ethylene glycol-bis[-aminoethyl ether]N,N,N',N'-tetraacetic acid) and probenecid (p-[dipropylsulfamoyl]benzoic acid) were purchased from Sigma/RBI (Natick, MA) and fura-2 acetoxymethyl ester and fluo-3 acetoxymethyl ester from Molecular Probes Inc. (Eugene, OR). Human orexin-A and -B were from Peninsula Laboratories (Merseyside, UK) or Neosystem (Strasbourg, France) and digitonin was from Merck (Darmstadt, Germany).

Peptide Synthesis. To avoid investigation of complex secondary structure changes caused by disulfide-bridge removals, we initially truncated orexin-A N-terminally to produce orexin-A_14-33, which is devoid of disulfide bonds, yet it has substantial activity on OX₁ and OX₂ receptors. A previous study has suggested that a significant part of the potency lost by truncation of orexin-A to orexin-A_15-33 is caused by the loss of the disulfide-bridges (Okumura et al., 2001). All the further mutagenesis was done on orexin-A_14-33. We both performed a further stepwise N-terminal truncation of this peptide and one-by-one replacement of amino acids with alanine ("alanine scan") (Table 1; Fig. 1). The truncated and alanine-scanned orexin-A peptides were synthesized using Fmoc (9-fluorenylmethoxycarbonyl) synthesis protocols with double or triple coupling reactions using TBTU (2-[H-benzotriazol-1-yl]-1,1,3,3-tetramethyluronium tetrafluoroborate) as the activator on a Symphony synthesizer (Rainin Instrument Co., Woburn, MA). Purifications were performed by reverse phase-HPLC on a -Pak C₁₈ column (15 µm; 100 Å; 25 × 100 mm) (Waters, Milford, MA) using a Waters liquid chromatography system consisting of a model 600 solvent delivery pump, a Rheodine injector, and an automated gradient controller (solvent A: H₂O-0.125% TFA [trifluoroacetic acid]; solvent B: CH₃CN-0.1% TFA; gradient: 15% B to 60% B in 20 min). Detection was carried out using a model M2487 variable wavelength UV detector connected to the Waters Millenium software control unit. The quality control was performed by analytical reverse phase-HPLC on a Waters -Pak C₁₈ (5 µm; 100 Å; 150 × 3.9 mm) column (same solvents as above; gradient: 0% B to 60% B in 20 min) using a Waters Alliance 2690 Separation Module equipped with a Waters 996 Photodiode Array Detector and by MALDI-TOF (matrix-assisted laser desorption ionization time-of-flight) mass spectrometry using a Voyager-DE instrument (Applied Biosystems, Foster City, CA).

View larger version (9K): [in this window] [in a new window]   Fig. 1.   Orexin-A and -B peptide sequences. The gray boxes mark the identical amino acids. The numbering of the amino acids is for orexin-A. Q* indicates the N-terminal pyroglutamoyl residue.

Media. The TES-buffered medium (TBM) consisted of 137 mM NaCl, 5 mM KCl, 1 mM CaCl₂, 1.2 mM MgCl₂, 0.44 mM KH₂PO₄, 4.2 mM NaHCO₃, 10 mM glucose, 1 mM probenecid, and 20 mM 2-([2-hydroxy-1,1-bis(hydroxymethyl)ethyl]amino) ethane sulfonic acid (TES) adjusted to pH 7.4 with NaOH.

Ca²⁺ Measurements. Both the fluorescent Ca²⁺ indicators fura-2 and fluo-3 were used to monitor changes in intracellular [Ca²⁺] ([Ca²⁺]_i) since they offer detection of low and high Ca²⁺ elevations, respectively, with high accuracy due to different K_d values for Ca²⁺ binding. The K_d of 224 nM was used for fura-2, whereas the K_d of 1000 nM was determined for fluo-3 in simultaneous measurement with fura-2. The low responses of the truncated C-terminal orexin-A-peptides were thus determined using fura-2, whereas all the other experiments were performed using fluo-3. It should be noted that the K_d value of the indicator does not affect the relative EC₅₀ values or the maximum responses determined. For the experiments, the cells were harvested using phosphate-buffered saline containing 0.2 g/l EDTA, loaded with fura-2 acetoxymethyl ester or fluo-3 acetoxymethyl ester (4 µM, 20 min, 37°C) in the culture medium (Ham's F-12) supplemented with 10 µg/ml bovine serum albumin and 1 mM probenecid and stored on ice as pellets (medium removed). For the measurement of intracellular free calcium, one pellet was resuspended in TBM at 37°C. The fluorescence was monitored in a stirred quartz microcuvette in a thermostated cell holder of either a Hitachi F-2000 or F-4000 fluorescence spectrophotometer at the wavelengths 340 or 340/380 (excitation), 505 (emission) for fura-2 or 480 nm (excitation), 540 nm (emission) for fluo-3. Experiments were calibrated by adding 60 µg/ml digitonin, which gives the maximum value of fluorescence, and 10 mM EGTA, which gives the minimum value of fluorescence. The leaked fura-2 and fluo-3 were measured in separate experiments by adding 10 mM EGTA, which chelates Ca²⁺ bound to the extracellular indicator. The corrected fluorescence values were used to calculate [Ca²⁺]_i.

Calculations and Data Analysis. The extracellular free [Ca²⁺] ([Ca²⁺]_e) was determined as described in Lund et al. (2000). Thus, addition of 1.5 mM EGTA in TBM gave a [Ca²⁺]_e of ~140 nM. Values are given as mean ± S.E. unless otherwise indicated; N refers to the number of batches of cells on which the measurements were performed. Nonlinear curve-fitting was performed using SigmaPlot for Windows 4.01 (SPSS Science, Chicago, IL). The difference in the potency (EC₅₀) and activity (maximum response) of orexin-A and orexin-B to elevate Ca²⁺ in 1 mM and 140 nM [Ca²⁺]_e (Table 2) was evaluated using Student's paired two-tailed t test. The differences in the activity (maximum response) between different peptides were evaluated using Student's nonpaired two-tailed t test.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Orexin-A and -B Differentially Activate OX₁ and OX₂ Receptors. As shown previously (Lund et al., 2000; Holmqvist et al., 2001; Kukkonen and Åkerman, 2001) orexin-A caused large elevations in intracellular [Ca²⁺] in OX₁ receptor-expressing CHO-K1 cells (Table 2). Orexin-B was 7-fold less potent (Tables 2 and 3; Fig. 2A). No difference in the potency between the ligands was seen for the OX₂ receptor (Tables 2 and 3; Fig. 2B). We have previously shown that OX₁-mediated Ca²⁺ response to low concentrations of orexin-A requires extracellular Ca²⁺ in CHO cells (Lund et al., 2000; Holmqvist et al., 2001; Kukkonen and Åkerman, 2001). This is not specific for orexin-A because an external Ca²⁺ dependence was also seen with respect to both orexin-A and -B response with both OX₁ and OX₂ receptors (Table 2; Fig. 2). Reduction in extracellular Ca²⁺ caused a considerable increase in the n_H value for all combinations of ligand/receptor. This change was largest for the effect of orexin-B on the OX₂ receptor.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Concentration-response curves for orexin-A and -B on OX1 (A) and OX2 (B) receptors in normal (1 mM) and low (140 nM) [Ca2+]e. N = 8. The fitted maximum response of each curve is normalized to the maximum response to orexin-A in 1 mM [Ca2+]e.

C-Terminal Orexin-A Peptides Are Agonistic for the OX₁ and OX₂ Receptors. C-Terminal truncation of orexin-A to orexin-A_14-33 (Fig. 1) increased the EC₅₀ to 27 nM from 4 nM (OX₁) or 7 nM (OX₂) (Tables 4 and 5). The maximum response, however, was not affected as calculated using the nonpaired t test. The effect of a further reduction in the peptide length is shown in Fig. 3. The response to 100 nM peptide fell when the peptide length was reduced to 18 amino acids (orexin-A_16-33) (Fig. 3,A and B). Further truncation to 15 amino acids (orexin-A_19-33) completely abolished the Ca²⁺ response to 100 nM peptide (Fig. 3, A and B). In a similar manner, both OX₁- and OX₂-expressing cells ceased responding to 1 µM peptide when its length was reduced to 12 amino acids (orexin-A_22-33), although the response to 10 µM peptide was retained (Fig. 3, A and B). With further truncation the fall in activity was eminent in the OX₁-expressing cells, and the C-terminal decapeptide (orexin-A_24-33) gave no response even at the concentration of 10 µM. The decline in activity with reduced peptide length was markedly lower with OX₂-receptors. At 10 µM, the nonapeptide (orexin-A_25-33) gave a robust response, and a slight response was still observed with the heptapeptide (orexin-A_27-33) but not with the hexapeptide (orexin-A_28-33) (Fig. 3, A and B).

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Ca2+ responses to different concentrations of truncated orexin-A peptides in 1 mM [Ca2+]e. Please observe the logarithmic ordinate. N = 3.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Graphic representation of the pEC50 values determined for alanine-scanned orexin-A14-33 peptides in 1 mM [Ca2+]e. The data are the same as presented in Tables 4 and 5.

View larger version (10K): [in this window] [in a new window]   Fig. 5.   Concentration-response curves for three different mutated orexin-A14-33 peptides in 1 mM [Ca2+]e. In A, the Leu20 to Ala mutation shifts the concentration response curve of orexin-A14-33 15- to 35-fold to the right. In B, the His21Ala mutation results in peptides with apparent highly cooperative response (steep concentration-response curves). Leu20Ala (A) and His26Ala (C) mutations show that the OX1 receptor is more sensitive to the orexin-A mutations than the OX2 receptor. N = 9.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Comparison of the pEC50 values of selected orexin-A14-33 peptides in normal (1 mM) and low (140 nM) [Ca2+]e. Please observe that the pEC50 of the Leu20Ala peptide could not be determined in 140 nM [Ca2+]e because of its low potency.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study we have investigated the differences between OX₁ and OX₂ receptors with respect to the specificity of activation by modified orexin peptides. The results obtained in this study indicate that both OX₁ and OX₂ receptors require very similar determinants from orexin-A to allow binding to and activation of the receptor, which indicates that analogous domains of OX₁ and OX₂ receptors interact with orexin-A. However, some interesting differences are seen, the most obvious of which being the less strict requirements of the OX₂ receptor for ligand binding.

As also previously shown, orexin-B had potency equal to orexin-A for the OX₂ receptor, but 10- to 100-fold lower relative potency for the OX₁ receptor. Thus, the OX₂ receptor cannot distinguish between orexin-A and -B as the OX₁ receptor does. The OX₁ receptor was much more affected by the N-terminal truncation of orexin-A than the OX₂ receptor. The first truncation of orexin-A to orexin-A_14-33, which removes the sulfhydryl bridges, essentially abolished the difference between OX₁ and OX₂ receptors with respect to orexin-A. Since orexin-B is N-terminally more or less truncated as compared with orexin-A, these data from the truncation of orexin-A may in part explain why orexin-B is as potent as orexin-A on the OX₂ receptor. However, domains interacting with the most N-terminal portion of orexin-A are not the only difference between the two receptors, as the OX₂ receptor also resisted further truncation far better than the OX₁ receptor. Alanine-scan identified residues of particular importance for orexin receptor activation. These residues were the same with respect to the activation of both OX₁ and OX₂ receptors, but the potency for the OX₂ receptor was less markedly affected. This further supports the notion of less strict requirements for binding and activation of the OX₂ receptor. Human OX₁ and OX₂ display 64% overall sequence identity (Sakurai et al., 1998). An even higher degree of identity and similarity are found between the proximal N termini and the extracellular loop 1 of the receptors. This might suggest that these parts of the receptors are important for the orexin binding, as is also suggested by the almost complete loss of orexin binding caused by the Glu54Lys (position homologous to Glu54 in human OX₂ receptor) mutation in the canine OX₂ receptor (Hungs et al., 2001). No obvious differences, which could explain the different behavior of OX₁ and OX₂ receptor, are seen.

All the peptides displayed slope factors above one and in addition, many of the truncated and alanine-scanned peptides, although not orexin-A itself, displayed slope factors (n_H) above 2.5. This may suggest cooperative binding, which in the case of G-protein-coupled receptors would indicate formation/presence of receptor di/oligomers. However, this has not been seen in binding studies (Sakurai et al., 1998; T. Holmqvist, K. E. O. Åkerman, J. P. Kukkonen, unpublished data), but the unsatisfactory conditions for the binding assay (see under Results) may mask this. Apparent cooperativity could also originate from the functional level. We have previously reported that orexin-A activates Ca²⁺ influx as a primary response mechanism. Ca²⁺ influx and an unknown signal from the receptor act synergistically to activate phospholipase C, since no phospholipase C activation is seen with Ca²⁺ influx alone, and in the absence of Ca²⁺ influx phospholipase C is activated only at 100× higher concentration of orexin-A (Lund et al., 2000; Kukkonen and Åkerman, 2001). We hypothesized that the reduced ability of some of the peptides to activate one of the Ca²⁺ responses, influx or release, could explain the high slope factors. Our experiments indeed confirmed that there are clear differences in the relative potencies between different mutant peptides toward these responses. The Ca²⁺ responses mediated by the OX₂ receptor were for most of the peptides less affected by removal of extracellular Ca²⁺ than the responses via OX₁ receptor. However, these differences do not explain the apparent cooperativity, and other explanations may be required.

It has been shown before that N-terminal truncation of orexin-A to orexin-A_15-33 (19 amino acids), which completely eliminates all cysteines, reduces the potency for the OX₁ receptor 60- to 170-fold (Darker et al., 2001; Okumura et al., 2001) but the reduction is only 20-fold for the OX₂ receptor (Okumura et al., 2001). In the present study the potency of orexin-A was decreased only 7- and 4-fold, respectively, for the OX₁ and OX₂ receptors upon truncation to orexin-A_14-33. Darker et al. (2001) have characterized the effect of further truncation of orexin-A on the activation of the OX₁ receptor: truncation of orexin-A to 19, 17, 16, 15, 10, and 5 amino acids progressively reduced the response to 10 µM peptide, and 5 and 10 amino acid-long peptides were inactive. In contrast to this, we did not observe any significant reduction in the maximum response until the chain length was reduced to 12 or 11 amino acids. In the present study, the shortened chain instead decreased the potency (increased the EC₅₀). This apparent discrepancy could be caused by a somewhat higher expression level in our CHO cells than those used in Darker et al. (2001). In our CHO cells the EC₅₀ value for orexin-A is >10-fold lower than the binding affinity for OX₁ receptors (T. Holmqvist, K. E. O. Åkerman, J. P. Kukkonen, unpublished data) as compared with only a 6-fold difference in Darker et al. (2001). Darker et al. (2001) have also shown that truncation of orexin-A to orexin-A_15-33 abolishes this difference in the EC₅₀ and binding affinity; thus, this truncation may decrease both the binding affinity and the efficacy of the peptide. Our semiquantitative estimation of the binding affinity suggests that further truncation may only reduce the binding affinity, and not the efficacy. In other words, once bound, both orexin-A_14-33 and the shorter peptides may have a similar ability to activate the receptor. The most C-terminal amino acids may be the most important for the activation of the receptor, whereas the additional N-terminal amino acids may increase the binding affinity and efficacy by making contact with the receptor and stabilizing orexin-A structure. The importance of the C terminus is evident since this area is highly conserved between orexin-A and orexin-B (Fig. 1). The present results support the view that for this kind of study, high receptor expression levels are useful.

Alanine-scan of residues 14 to 26 in orexin-A_14-33 identified three areas of interest, which were the same in both OX₁ and OX₂ receptors: amino acids 15 to 17, 20, and 25 to 26. Alanine-scan of orexin-A_15-33 was performed in Darker et al. (2001) with results similar to the results in the present study with respect to OX₁ receptor activation. This study also showed that the receptor response is extremely sensitive to mutations in the outermost C terminus (amino acids 26-33), as can be expected from the highly similar C termini of orexin-A and -B. In contrast to Darker et al. (2001), we failed to detect any decrease in the potency by mutation of residues 21 to 24. Neither we nor Darker et al. (2001) detected any mutation that would markedly increase the potency of the peptide. The most interesting mutation is the Leu20Ala, which causes an over 10-fold drop in the potency of the orexin-A_14-33. In contrast to the findings of Darker et al. (2001) we saw almost no effect of the mutation of the adjacent amino acids, Leu19 or His21. Mutation of a leucine to alanine does not cause any change in the charge of the peptide; neither can the effect be explained by steric hindrance, which should instead be diminished by the less bulky alanine. Therefore, the effect of this mutation is likely to affect the secondary or tertiary structure of orexin-A_14-33 rather than the interaction with the orexin-binding site. If orexin-A, as orexin-B (Lee et al., 1999) forms an -helical structure in this region, then replacement of leucine with alanine should also not have any remarkable effect on the secondary structure. Since the three-dimensional structure of orexin-A is unknown, no conclusion of its structure and the effects of mutations thereon can be made. The alanine-scanned peptides activated the OX₂ receptor with potencies very similar to those for the OX₁ receptor. However, the OX₂ receptor once again proved less "fussy" in its requirements on the orexin-A_15-33 peptide than the OX₁ receptor.

Orexin receptor subtypes are somewhat differentially expressed in the CNS and in the periphery, but the significance of this is unknown since no separate cellular functions have been ascribed to the receptor subtypes, and the relative amounts of orexin-A versus orexin-B seem rather constant in the different CNS areas (Kukkonen et al., 2002). Therefore, some of the major issues are whether orexin receptor subtypes can couple to different signaling pathways and whether different orexin peptides can cause "signal-trafficking" (Kenakin, 1995) even via a single receptor subtype. The data from the experiments with reduced extracellular [Ca²⁺] suggest that orexin peptides may have an inherent ability to do this. Also, some in vivo and in vitro data suggest that orexin-B is more efficacious than orexin-A (see Kukkonen et al., 2002), which is unexplainable based on the recombinant pharmacology. We propose that the physiological roles of orexins can only be elucidated in the light of much further data on the receptor subtype specificity in ligand binding and signaling pathway activation.