A Preformed, Ordered Structure of a 25-residue Peptide Derived from a Major Histocompatibility Complex Class I Antigen Is Required to Affect Insulin Receptor Function*

It was recently shown that a 25-residue peptide, Dk-(61-85), derived from t h e a l domain of a murine major histocompatibility class I molecule (H-2Dk), affects in- sulin receptor functions (Hansen, T., Stagsted, J., Ped-ersen, A., Goldstein, A., and L. Proc. Natl. Acad. Sci. A. 86,3123-3126; M., Hansen, T., Goldstein, A., and Olsson, L. (1990) Cell 62, 297-307). We now report that this peptide can reversibly assume a biologically active or inactive state as measured in the rat adipocyte glucose uptake assay, implying that the peptide has at least two interconvertible conformations. The peptide has an ordered conformation in 0.1 M HC1 or 0.1 M NaCl stock solution as shown by circular dichroism, but has a disordered molecular structure and is inactive when dissolved in H20. The biologically active peptide forms liquid crystals at the stock solu- tion concentration (1 mM), so the CD spectra do not provide information on the secondary structure. Under

It was recently shown that a 25-residue peptide, Dk-(61-85), derived from t h e a l domain of a murine major histocompatibility class I molecule (H-2Dk), affects insulin receptor functions (Hansen, T., Stagsted, J., Pedersen, L., Roth, R. A., Goldstein, A., and Olsson, L. (1989) Proc. Natl. Acad. Sci. U. S. A . 86,3123-3126; Stagsted, J., Reaven, G. M., Hansen, T., Goldstein, A., and Olsson, L. (1990) Cell 62, 297-307). We now report that this peptide can reversibly assume a biologically active or inactive state as measured in the rat adipocyte glucose uptake assay, implying that the peptide has at least two interconvertible conformations. The peptide has an ordered conformation in 0.1 M HC1 or 0.1 M NaCl stock solution as shown by circular dichroism, but has a disordered molecular structure and is inactive when dissolved in H20. The biologically active peptide forms liquid crystals at the stock solution concentration (1 mM), so the CD spectra do not provide information on the secondary structure. Under all conditions tested, biological activity (measured after transfer to assay buffer) is associated with an ordered conformation in stock solution. Biological activity and an ordered conformation of the peptide in HzO stock solution can be induced by increasing ionic strength (>lo0 mM NaCl for maximal effect) or increasing pH (>5 for maximal effect). The induction rate of the ordered conformation is slow with a halfmaximal value obtained after -20 min. Both biological activity and the ordered structure are lost upon heating of stock solution to 90 "C or upon transfer to assay buffer. A similar correlation of ordered structure with biological activity was observed with two truncated peptides derived from Dk-(61-85). It is inferred from these results that the Dk-(61-85) peptide and related peptides only affect insulin-stimulated glucose uptake in rat adipocytes if they have assumed an ordered conformation in stock solution prior to transfer to assay buffer and exposure to cells.
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Biological activity of some peptides can be potentiated by amino acid modifications that enhance their amphiphilicity and that often also results in an a-helical secondary structure (3)(4)(5). In line with this, a correlation between biological activity of peptides and their solvent conditions has been demonstrated (6,7). We recently reported that a 25-residue peptide, Dk-(61-85), derived from the a1 domain of a murine major histocompatibility complex (MHC)' class I antigen, affects key insulin receptor functions, including tyrosine kinase activity, insulin-stimulated glucose uptake, and insulin-induced receptor internalization (1,2). We now report that the peptide, which has an amino acid sequence compatible with an amphiphilic secondary structure, can occur both in a biologically active and in an inactive state. The active form can be reversed to an inactive form by changing the solvent of the peptide stock solution from HCl or NaCl to H20. Circular dichroism studies show that the active peptide in NaCl stock solution has an ordered, complex structure, whereas the inactive peptide in H 2 0 has a spectrum consistent with a disordered structure. Further, and as a novel observation, the ordered conformation of this amphiphilic peptide must be assumed in the stock solution prior to transfer to assay buffer and exposure to cells. Thus, this peptide differs from peptides with membrane affinity, which only assume a secondary structure upon interaction with cell membranes (3). It is concluded that the Dk-(61-85) peptide may assume different, but interconvertible conformations, that differ in biological activity and that the biological potency might be enhanced by modifications that stabilize the active conformation(s).

MATERIALS AND METHODS
Peptides-The sequence of the peptide designated Dk-(G1-85) was derived from the a1 domain of the heavy chain of murine MHC class I (H-2k). Other peptides used in this study are Dk-(61-84) and DL-(69-85). The sequence of Dk-(G1-85) is shown in Fig. 1. The biological activity of the peptide was described previously (1, 2). All peptides were synthesized by Applied Biosystems, Inc. (Foster City, CA). The crude peptides were purified by preparative HPLC. Each peptide was more than 95% pure, as judged by analytical HPLC monitoring of absorbance at 214 and 278 nm (Fig. 1). Identities of the peptides were confirmed by amino acid composition and mass spectrometry.
After purification, peptide was lyophilized overnight from CH,CN/ trifluoroacetic acid and then solubilized at 25 "C in one of two stock solvents at 1 mM:O.l M HC1 (Pierce Chemical co., Sequanal grade) or H20 (Baxter, Burdick & Jackson Division, HPLC grade), as illustrated in Fig. 1 (step I). Sodium chloride (final concentration, 0.1 M) was added in some experiments to the stock solution of watersolubilized peptide (Fig. 1, step 2 ) . In step 3 (incubation in assay buffer without cells), incubation time was varied from 0 to 180 min.
Glucose Uptake by Adipocytes-Adipocytes were prepared as described (2). Briefly, minced epididymal fat pads were incubated in KRHB with 5 mM D-glucose and 1 mg/ml collagenase (type I, Worthington) and digested for 1 h at 37 "C. The adipocytes were washed five times (the conventional three washes result in substantial peptide degradation in KRHB), with each wash performed in 10 times the Time (rnin)  The purity of each peptide was >95% as judged by absorbance at 214 and 278 nm. Right, flow diagram for peptide application in the glucose uptake and CD assays (see also "Materials and Methods").
cell volume, and diluted to a 10% (v/v) suspension (-lo6 cells/ml). Fifty pl of adipocyte cell suspension was pipetted into Minisorp tubes (Nunc, Denmark) and equilibrated at 37 "C with gentle shaking for 30 min. Insulin and peptides were added in 50 p1 of KRHB and incubated for 30 min at 37 "C. Then ['4C]glucose tracer (-100,000 dpm) was added, and the incubation was continued for another 30 min at 37 "C. The cells were harvested by centrifugation on top of silicone oil, and the cell-associated radioactivity was determined by scintillation counting. Circular Dichroism Spectroscopy-CD spectra were recorded on a JASCO 5-600 calibrated against d-camphorsulfonic acid using A t (290.5 nm) = +2.38 M" cm". Cylindrical cells with path lengths of 0.01 cm were used for recording spectra of 1 mM peptide stock solutions, and path lengths of 0.10 cm were used for 30 p M peptide in KRH. The temperature was varied within a range of 30-90 "C and controlled to 0.1 "C using a refrigerated recirculating waterbath. Fig. 2A shows that, when diluted in KRHB to a final concentration of 30 PM, the peptide from the stock solution in H 2 0 had no effect on the insulin-stimulated glucose uptake by rat adipocytes, whereas peptide dissolved in 0.1 M HCl in the stock solution enhanced insulin-stimulated glucose uptake. The difference between biologically active and inactive peptide was significant (p < 0.001, t test). Further, the stock solvent effects on biological activity of the peptide were fully reversible as demonstrated by the activity of HzOor HC1-solubilized peptides upon lyophilization and subsequent resolubilization in HC1 or HzO ( Fig. 2A). It was later found that the biological activity could be induced in the inactive peptide from the HzO stock solution by addition of NaCl to a final concentration of 0.1 M at least 0.5 h prior to transfer to assay buffer. Therefore, biologically active peptide was, for most of the following experiments, obtained from a H,O-solubilized peptide stock solution to which NaCl was added (NaC1 stock solution). Fig. 2B shows that the biological activity of the peptide from NaCl stock solution had an ECso -12 PM and that the peptide from HzO stock solution was about 20-fold less potent. Peptide concentrations higher than 300 PM could not be tested, as precipitation became significant, and the exact difference in biological potency between peptide in NaCl and HzO stock solutions was therefore not possible to estimate.

Effect of Soluents on Peptide Actiuity-
HPLC analysis on a reversed-phase CIS column with a gradient of 20-40% CH&N in 5 mM trifluoroacetic acid over 5 min gave almost identical retention times (mean k S.E., n = 4) for stock solution of the peptide in HCl (2.96 k 0.03 min) and HzO (3.00 f 0.02 min), with no indication of twin peaks in several coelution runs (data not shown). HPLC analysis of supernatants from adipocytes incubated with 3Hlabeled peptide at 30 PM for 30 min in assay buffer showed -20% degradation (data not shown), irrespective of whether the stock solution of 3H-labeled peptide had been prepared in HC1 or HzO.
Circular Dichroism Analysis of the Peptide and Truncated Deriuatiues-The CD spectrum for the peptide in HzO indicates a random coil conformation, whereas the spectrum for the peptide in NaCl implies a complex, ordered structure (Fig.  3). The CD spectrum for the peptide in NaCl was independent of the distance between cuvette and detector, thus excluding chiral scattering artifacts (data not shown). The CD analysis of two truncated derivatives of the peptide, Dk-(69-85) and to circular dichroism ( A € ) per amino acid residue ( A € = @222nm/3298). 100% ahelix has A6 = -9.6 at 222 nm (9). It may be noted that none of these spectra have a local minimum at 222 nm (Dk-(61-85) in NaCl has local minima at 215 and at 198 nm; Dk-(61-85) in HzO has one at 195 nm). The rate of the CD signal change at 220 nm after addition of NaCl to 0.9 mM Dk-(61-85) stock solution in Hz0 is shown in the inset.

Conformations and Biological Activity
Dk-(61-84), demonstrated, for Dk-(69-85) (biologically active), a pattern similar to the peptide in NaCl and, for Dk-(61-84) (biologically inactive), a random coil spectrum, although shifted to the right. The inset shows the rate at which the CD signal at 220 nm changes from a random coil spectrum to a spectrum for a complex ordered structure after addition of NaCl to the peptide stock solution in H20.
Peptide stock solution (0.9 mM) in NaCl was found to be birefringent with characteristics of liquid crystals when observed between crossed Polaroid filters, whereas peptide stock solution in HzO was nonrefringent. Biological Activity and CD Signal as Related to Salt Concentration, pH, Temperature, and Time in Assay Buffer-Biological activity and an ordered peptide conformation could be induced in HzO stock solutions of the peptide by an increase in ionic strength (EC,, -50 mM NaC1) or p H (apparent pK, -4.2) with the curves for the amount of peptide in a n ordered conformation as measured by CD and those for biological activity being generally similar (Fig. 4, A and B). The temperature dependency for melting of the peptide conformation and disappearance of biological activity in stock solutions of peptide in NaCl are seen in Fig. 4C with a T,,, -70 "C. to the adipocytes at 30 N M for the standard 60-min glucose uptake assay. For CD measurements, the peptide was diluted to 30 PM in KRH, and the CD signal at 220 nm was recorded within 1 min after dilution. Each point for both biological activity (open symbols) and CD signal (filled symbols) represents mean & S.E. of three experiments, where the maximum activity is set to 100%. B, 1 mM Dk-(61-85) in HzO (pH = 3.0) was titrated to the indicated pH using 0.1 M NaOH or HCl and subsequently incubated at 25 "C for 90 min. Biological activity and CD measurements were performed as in A . The data points (mean of triplicate samples) for both biological activity and CD signal are from three separate experiments, where maximal biological or CD signal was set to 100% in each experiment. The highest ionic strength (-4 mM final) contributed by the added base or acid was insufficient to induce activity by ionic strength alone (see A ) . C, Dk-(61-85) in NaCl at a concentration of 0.9 mM was heated from 30 to 90 "C, and samples were taken out at various temperatures, transferred to 37 "C assay buffer, and tested in the 60-min glucose uptake assay. The results are means & S.E. of three experiments. The corresponding CD signal was measured at 0.5 "C intervals and is the mean & S.E. of four experiments, where the CD signal at 30 "C is set to 100%. D, decay of biological activity and CD signal at 220 nm of Dk-(61-85) peptide in NaCl stock solution upon transfer to assay buffer. The x axis indicates the time after transfer to KRH but prior to application in the glucose uptake assay, i.e. the duration of step 3 in Fig. 1 between the biologically active peptide and large peptide aggregates.

DISCUSSION
Previous studies have shown that the peptide affects key insulin receptor functions, including tyrosine kinase activity, insulin-stimulated glucose uptake, and insulin-induced receptor internalization (1,2). In the present study, biological activity was determined with the rat adipocyte glucose uptake assay. The biological activity of the peptide from NaCl stock solution compared to that from H20 stock solution shows that the peptide can occur in at least two different states, of which one is biologically active, the other inactive. Conversion from one form to the other is reversible and is affected by the ionic strength, pH, and temperature of the solvent. CD analysis reveals a n ordered conformation in those peptide stock solutions that are biologically active upon transfer to KRHB (Table I). The stock solution of the biologically active Dk-  A typical a-helical CD spectrum shows local minima at 222 and 206 nm (8), but the spectrum for the peptide in NaCl has local minima that are both shifted leftward -5-10 nm. This suggests a complex structure that may contain both a-helix and p-sheet conformations. However, under these conditions the biologically active peptide forms liquid crystals, and the CD spectra can therefore not be used to deduce any information on the secondary structure of the individual peptide molecules. The slow rate of induction (minutes) of the ordered structure after addition of NaCl to the peptide in HzO suggests that the biologically active conformation requires interaction between peptide molecules and is consistent with the formation of liquid crystals. In contrast, intramolecular conformational changes are known to occur in milliseconds.
The linear sequence is not sufficient for the biological effect of the peptide. The peptide must not only acquire a certain conformation in order to be biologically active, but it is also evident that this active conformation must exist in the stock solution prior to addition to the assay buffer. Both the complex, ordered structure and biological activity disappear if the active peptide is incubated for 20-30 min in cell-free KRH prior to the glucose uptake assay.
Peptide sequence alterations that result in increased amphiphilicity and a-helix formation also often result in enhanced biological activity of several hormones such as calcitonin (9-11) and p-endorphin (12)(13)(14). In addition, the assumed association between amphiphilic a-helix formation and biological activity has been explored successfully to design biologically active peptide analogues of the honey bee venom melittin (15). Based on the correlation between amphiphilicity and biological activity, Kaiser's group suggested that membrane affinity is an intrinsic property of a peptide and determined by the amino acid sequence, which on the other hand defines its ability to assume an amphipathic structure. The Dk-(61-85) peptide must assume its active conformation prior to interaction with the adipocytes, and it is consequently different from peptides with membrane affinity, as defined by Kaiser and Kezdy (3). However, it is conceivable that peptides with an ordered structure create an amphiphilic environment when they aggregate, promoting further recruitment of peptides in an ordered amphiphilic structure into the aggregate. The correlation between biological activity and the occurrence of a complex structure of the peptide in stock solution further suggests that peptide molecules in the biologically active conformation are contained in these complex structures. The aggregate may in that case serve as a reservoir for active monomeric or oligomeric peptides. It may also be noted that small differences in the linear sequence of peptides, which affect their amphiphilicity, can have a major impact on their binding to MHC class I and class I1 antigens and, thereby, their immunogenicity (16,17). Our observations described above indicate, however, that conformational differences of a peptide caused by different solvent conditions may result in variable immunogenicity, which therefore is not solely determined by the amino acid sequence.
Enhancement of the potency of active peptide is desirable for several reasons, including the fact that low potency has thus far excluded binding studies to determine binding site(s) and affinities to insulin receptor and whole cells. Based on the observations above, we suggest that enhancement of potency may be achieved by peptide modifications that promote and stabilize the ordered conformation associated with biological activity.