Development of specific and non-specific somatostatin analogs

Biological activity of six somatostatin analogs has been investigated. In these analogs, disulfide bond is replaced by ethylene bond cyclized with alpha-amino suberic acid. In addition, they contain unique D-configuration in both Trp8 and Cys14 moiety with dicarba substitution. An analog of the short chain length, C omega 7-cyclo (Phe6-Phe7-D-Trp8-Lys9-Thr10-Phe11-D-Asu14) (analog 4) has suppressive effect for GH, but not for other hormones. Analog 6, C omega 9-cyclo(Asn5-Phe6-Phe7-D-Trp8-Lys9-Thr10-Ph e11-Thr12-D-Asu14), has suppressed GH and insulin secretion, but not for gastrin and glucagon. Analog 1, C omega 11-cyclo (Lys4-Asn5-Phe6-Phe7-D-Trp8-Lys9-Thr10-Phe11- Thr12-Ser13-D-Asu14] and 5, C omega 9-cyclo (Lys4-Asn5-Phe6-Phe7-D-Trp8-Lys9-Thr10-Phe11-D-+ ++Asu14) have broad suppressive effect for GH, gastrin, insulin and glucagon release after arginine infusion. The shortest analog, analog 2, C omega 5-cyclo (Phe7-D-Trp8-Lys9-Thr10-D-Asu14) has weak suppressive effect of GH, insulin and glucagon secretion, and it is suggested that Phe6 and Phe11 are necessary for the appearance of suppressive effect of GH. Specific analog, analog 4, may be useful for the future treatment for acromegaly and diabetic retinopathy. Nonspecific analogs, 1 and 5 are candidates for the clinical application of wide variety.     

Proton magnetic resonance spectra of adrenodoxin: features of the aromatic region

This paper presents the first 1H-NMR spectra of the aromatic region of adrenodoxin, a mammalian mitochondrial 2Fe-2S non-heme iron ferredoxin. One-dimensional proton NMR spectra of both reduced and oxidized adrenodoxin were recorded as a function of pH. Resonances due to two of the three histidines of adrenodoxin gave sharp signals in the one-dimensional proton NMR spectra. The pKa values of the resolved histidine resonances in the oxidized protein were 6.64 +/- 0.03 and 6.12 +/- 0.06. These values were unchanged when adrenodoxin was reduced by the addition of sodium dithionite. In addition, the oxidized protein showed a broadened histidine C-2H resonance with a pKa value of approx. 7. This resonance was not apparent in the spectra of the reduced protein. The resonances due to the single tyrosine in adrenodoxin were identified using convolution difference spectroscopy. In addition, a two-dimensional Fourier-transform double quantum filtered (proton, proton) chemical shift correlated (DQF-COSY) spectrum of oxidized adrenodoxin was obtained. The cross peaks of the resonances due to the tyrosine, the four phenylalanines, and two of the three histidines of adrenodoxin were resolved in the DQF-COSY spectrum. Reduction of the protein caused several changes in the aromatic region of the NMR spectra. The resonances assigned to the C2 proton of the histidine with a pKa of 6.6 shifted upfield approx. 0.15 ppm. In addition, when the protein was reduced one of the resonances assigned to a phenylalanine residue with a chemical shift of 7.50 ppm appeared to move downfield to 7.82 ppm.(ABSTRACT TRUNCATED AT 250 WORDS)     

Conformational analyses by 1H NMR and computer simulations of cyclic hexapeptides related to somatostatin containing acidic and basic peptoid residues.

We report the conformational analysis by 1H NMR in DMSO and computer simulations involving distance geometry and molecular dynamics simulations at 300K of peptoid analogs of the cyclic hexapeptide c-[Phe11-Pro6-Phe7-D-Trp8-Lys9-Thr10]. The analogs c-[Phe11-Nasp6-Phe7-D-Trp8-Lys9-Thr10](1), c-[Phe11-Ndab6Phe7-D-Trp8-Lys9-Thr10] (2) and c-[Phen11-Nlys6-Phe7-D-Trp8-Lys9-Thr10](3) where Nasp denotes N-(2-carboxyethyl) glycine, Ndab N-(2-aminoethyl) glycine and Nlys N-(4-aminobutyl) glycine are subject to conformational studies. The results of free and restrained molecular dynamics simulations at 300K are reported and give insight into the conformational behaviour of these analogs. The compounds show two sets of nuclear magnetic resonance signals corresponding to the cis and trans orientations of the peptide bond between residues 11 and 6. The backbone conformation of the cis isomers that we believe are the bioactive isomers of the three compounds are very similar to each other while there are larger variations amongst the trans isomers. The binding data to the isolated receptors show that the introduction of the Nlys residue in analog 3 leads to an enhancement of binding potency to the hsst5 receptor compared with analog 2 while maintaining identical binding potency to the hsst2 receptor. The Nasp6 analog 1 binds weakly to the hsst2 and is essentially inactive towards the other receptors. Comparison of the conformations and binding activities of these three analogs indicates that the Nlys residue extends sufficiently far to allow binding to a negatively charged binding domain on the hsst5 receptor. According to this model, the Ndab analog 2 cannot extend far enough to allow for binding to the receptor pocket. The loss of activity observed for the Nasp6 compound 1 indicates that the presence of a negatively charged residue in position 6 is unfavorable for binding to the hsst receptors.     

The degradation of somatostatin by synaptic membrane of rat hippocampus is initiated by endopeptidase-24.11

Somatostatin was degraded by the synaptic membrane from rat hippocampus. Cleavage products were separated by reversed phase high performance liquid chromatography and identified by amino acid composition analyses and N-terminal amino acid and sequence determinations around the cleavage sites. Fragments produced from the cleavages at both or either sites between the Phe6-Phe7 and/or between the Thr10-Phe11, together with free phenylalanine and tryptophan, were major cleavage products, followed by that produced from the cleavage of the Asn5-Phe6 bond. The accumulation of the major cleavage products, as well as the initial cleavage of somatostatin, was strongly inhibited by metal chelators and also by specific inhibitors of endopeptidase-24.11 (EC 3.4.24.11), phosphoramidon and thiorphan. The inhibitor susceptibility of the synaptic membrane toward somatostatin was similar to that toward Leu-enkephalin, a natural substrate of endopeptidase-24.11. Furthermore, endopeptidase-24.11 purified from rat brain hydrolyzed somatostatin at the cleavage sites identical to those by the hippocampal synaptic membrane. Thus, it can be concluded that endopeptidase-24.11 plays a major role in the initial stage of somatostatin degradation in rat hippocampus.     

Carbon-13 NMR investigations on ribonuclease A.

The proposed interaction between the amino acid residues Asp 14 and His 48 of ribonuclease A has been confirmed by 13C-NMR spectroscopy. The titration behaviour of the resonance of the side-chain carboxyl group of Asp 14 suggests a pKa of 6.5--7.0 for His 48. An equilibrium between different conformation process of His 48. Upon this deprotonation a hydrogen bond between the side-chains of Asp 14 or His 48 and Tyr 25 seems to be formed as is suggested by the behaviour of a tyrosine C zeta resonance assigned to Tyr 25. One phenylalanine resonance broadens and moves upfield on the addition of the inhibitor Cyd-2'-P, being therefore assigned to Phe 120. The behaviour of this resonance suggests that the upfield shift results from the anisotropy of the cytidine ring. Three signals are assigned to the three Phe residues.     

Peptide conformations--49(1): synthesis and structure-activity relationships of side chain modified peptides of cyclo(-D-Pro-Phe-Thr-Lys-Trp-Phe.)

Cyclic hexapeptide analogues representing the modified retro sequence of the amino acid residues 7-11 of natural somatostatin are known to protect liver cells from phalloidin poisoning. To determine the influence of steric, lipophilic, and charge effects on (a) the conformation of the backbone and the aromatic side chains and (b) the biological response, the side chains of Phe2, Lys4, and Phe6 of cyclo(-D-Pro1-Phe2-Thr3-Lys(Z)4-Trp5-Phe6-), 1a, one of the most active peptides found so far, were modified by various residues. The discussion of conformationally relevant parameters proves that neither backbone conformations nor populations of aromatic side chain rotamers were altered by these substitutions. The potency of these derivatives in a cytoprotection assay varies by at most one order of magnitude (more or less active than the parent peptide 1a). A qualitative evaluation of lipophilic, steric, and charge effects reveals the dominance of lipophilic effects of aromatic residues; the most potent compounds contain aromatic substructures in the side chain of Lys4.     

Purification of the main somatostatin-degrading proteases from rat and pig brains, their action on other neuropeptides, and their identification as endopeptidases 24.15 and 24.16.

The main somatostatin-degrading proteases were purified from rat and pig brain homogenates and characterized as thiol- and metal-dependent endoproteases. Two types of proteases with apparent native and subunit molecular masses of 70 kDa and 68 kDa could be differentiated in both species. Beside somatostatin, both hydrolyzed several other neuropeptides with chain lengths between 8 and 30 amino acid residues. Cleavage sites were generally similar or identical, but some clear exceptions were observed for enzymes from both species which could be used to differentiate between the two proteases. The 68-kDa protease cleaved somatostatin at three bonds (Asn5-Phe6, Phe6-Phe7 and Thr10-Phe11) and neurotensin only at the Arg8-Arg9 bond, whereas the 70-kDa protease digested somatostatin at only two bonds (Phe6-Phe7 and Thr10-Phe11) and neurotensin as well as acetylneurotensin-(8-13) additionally (pig protease) or almost exclusively (rat protease) at the Pro10-Tyr11 bond. Relative rates for the digestions of various peptides were, however, more dependent on the species than on the type of protease. Cleavage sites for angiotensin II, bradykinin, dynorphin, gonadoliberin and substance P were, apart from different rates, identical for both proteases. In both species the 68-kDa protease was found to be mainly, but not exclusively, soluble and not membrane-associated, whereas the inverse was detected for the 70-kDa protease. Based on distinct molecular and catalytic properties, the 68-kDa protease is supposed to be congruent with the endopeptidase 24.15 (EC 3.4.24.15), the 70-kDa protease with endopeptidase 24.16 (EC 3.4.24.16, neurotensin-degrading endopeptidase). This investigation demonstrates that both proteases hydrolyze various neuropeptides with similar cleavage sites, but with species-dependent activity. Species-independent distinctions are the exclusive action of endopeptidase 24.16 on acetylneurotensin-(8-13) and liberation of free Phe from somatostatin only by endopeptidase 24.15.     

Hydrolysis of the synthetic chromophoric hexapeptide Leu-Ser-Phe(NO2)-Nle-Ala-Leu-OMe catalyzed by bovine pepsin A. Dependence on pH and effect of enzyme phosphorylation level

Inactivation of substance P and its C-terminal hexapeptide analog [p-Glu6]substance P6-11 was studied in rat parotid and hypothalamic slices. It was found that in the parotid slice system the decay of substance P induced K+ release occurs concurrently with a decrease in the biologically active concentration of the peptide in the medium. The inactivation was further studied using [p-Glu6]substance P6-11 as substrate in the parotid and in the hypothalamic slice systems. In both tissue preparations the hexapeptide is degraded to small peptide fragments by metalloendopeptidase. Separation of the peptide fragments by high performance liquid chromatography and determination of their amino acid composition showed that in the hypothalamic slice system the major cleavage of the hexapeptide analog occurs between Phe8-Gly9 with minor cleavage sites between Phe7-Phe8 and Gly9-Leu10. In the rat parotid slice system the major cleavage occurs between Gly9-Leu10 with a minor cleavage site between Phe7-Phe8. The degradation of the hexapeptide analog in the hypothalamic system was inhibited 77% and 67% by treatment with 1 mM p-chloromercuriphenylsulfonate and p-chloromercuribenzoate, respectively, whereas in the parotid system these reagents inhibited the degradation of the hexapeptide only by 15% and 8%. These results may indicate that different proteases in the parotid and hypothalamus are involved in degradation of substance P. Kinetic studies, including the use of various inhibitors as well as competition by the peptide hormones somatostatin, LHRH, TRH and Leu-enkephalin-NH2, revealed that in both tissues the hexapeptide analog is a preferred substrate for degradation by protease of considerable specificity towards the C-terminal sequence of substance P. It is suggested that this metalloendopeptidase may be important in the termination of the substance P response.     

Nuclear-magnetic-resonance studies of ferrocytochrome c. pH and temperature dependence

The pH dependence and the temperature dependence of the nuclear magnetic resonance spectrum of horse ferrocytochrome c are described. This protein is very stable; it maintains an ordered structure over the pH range 4 to 12 at 25 degrees C and over the temperature range 4 degrees C to 97 degrees C at pH 7.0. The dynamic characteristics of the conformation of ferrocytochrome c were investigated. Particular emphasis was laid on the aromatic resonances and resonances of methyl groups shifted far upfield. Tyr-48 and Phe-46 were found to be relatively immobile whilst a region of the protein close to Ile-57 was found to be relatively flexible.     

Proteins from the prokaryotic nucleoid: 1H NMR study of the quaternary structure of Escherichia coli DNA binding protein NS (HU)

The quaternary interactions of Escherichia coli DNA binding proteins NS1, NS2, and NS (NS1 + NS2) have been studied by 1H NMR spectroscopy at 400 MHz following the reversible spectral changes produced by temperature increases on the resonances (Phe ring and His C-2 protons) whose spectral characteristics reflect the formation and dissociation of either homologous or heterologous interactions. These changes include (a) a progressive intensity decrease of the Phe resonances shifted to high field by stacking interactions, (b) a progressive intensity increase of the resonances due to freely rotating Phe, and (c) splitting of the His C-2 proton resonance. The association constants and thermodynamic parameters for the homologous and heterologous interactions were calculated from the molar fractions of the relevant molecular species by assuming that the above effects are due to the existence of simple association equilibria. It was found that two (out of three) phenylalanine residues of each polypeptide chain are involved in quaternary interactions. Quantitative data concerning the internal mobility and mutual orientations in aggregates of these Phe rings were also obtained. From the calculated association constants, from comparison of these data with recent protein-protein cross-linking results [Losso, M. A., Pawlik, R. T., Canonaco, M. A., & Gualerzi, C. O. (1986) Eur. J. Biochem. 155, 27-32], and from other considerations, we suggest that even though stacking of the Phe rings occurs at the interface between monomers, the temperature-dependent alteration of the Phe spectrum monitors shifts of the dimer in equilibrium tetramer equilibrium whereas the splitting of the His C-2 proton resonance most likely monitors the equilibrium between tetramers and larger aggregates.     

Assignment of resonances in the phosphorus-31 nuclear magnetic resonance spectrum of poly[d(A-T)] from phosphorothioate substitution

Two phosphorothioate analogues of poly[d(A$-T)] have been synthesized enzymatically. In one, poly[d(A$-T)], dTMP is replaced by thymidine 5'-O-phosphorothioate; in the other, poly[d(T$-A)], dAMP is replaced by 2'-deoxyadenosine 5'-O-phosphorothioate. The 31P NMR spectrum of poly[d-(A-T)] in solutions at low salt concentration shows two resonances at 51.80 and -4.25 ppm relative to trimethyl phosphate. The corresponding values for poly[d(T$-A)] are 51.51 and -4.43 ppm. These data allow the assignment of the downfield resonance at -4.23 ppm in poly[d(A-T)] to the phosphate group of d(TpA) and the resonance at -4.41 ppm to that of d(ApT). Thus, strong evidence is provided for a repeating dinucleotide structure. A comparison of the 31P NMR spectra of the various polymers in solutions of 2 M CsF reveals that both resonances are shifted upfield by approximately 0.9 ppm in the case of the phosphorothioates and by 0.2 or 0.4 ppm in the case of the phosphates. An upfield shift of about 0.18 ppm can also be observed for the two corresponding dinucleoside monophosphates. Thus, the upfield shift induced by high concentrations of CsF is not specific for the polymer backbone.     

Fluorine nuclear magnetic resonance spectra of rabbit carbonmonoxyhemoglobin

Carbonmonoxyhemoglobin prepared from protein isolated from rabbits maintained on a diet supplemented with 4-fluorophenylalanine (Phe (4F)) has been studied by fluorine NMR spectroscopy. Substitution of Phe(4F) appears to take place randomly at the sixteen nonequivalent phenylalanine positions of the globins; examination of hybrid hemoglobins in which only one type of globin chain contained the fluorinated amino acid, as well as changes in the spectrum upon exposure to oxygen, aided in the assignment of fluorine resonances from the alpha- or beta-globin chains. The effects of modification of Cys-beta-93 with a spin label and variation of pH and sample temperature on the spectrum were also examined. These data in association with theoretical estimates of aromatic ring current and van der Waals effects on chemical shifts were used to support tentative assignments of several signals observed to specific amino acid residues. Evidence suggesting the presence of two conformational forms of the protein, possibly due to disorder of the heme groups, is described.     

Solution conformation of enkephalin. A nuclear magnetic resonance study of 13C-enriched carbonyl carbons in [Leu5]-enkephalin.

By using 13C enrichment in [Leu5]-enkephalin, it has been possible to improve the assignment of carbonyl resonances in the nuclear resonance spectrum and to remove some of the ambiguities in the derived phi and chi dihedral angles, thereby providing information about the conformation of this molecule in solution. The combined use of 13C and 1H nuclear magnetic resonance experiments leads to the conclusion that [Leu5]0enkephalin contains a type I beta bend at residues Gly3-Phe4 in dimethyl-d6 sulfoxide (Me2SO0d6) solution. Furthermore, the side chains of Tyr1, Phe4, and Leu5 exist predominantly in one conformation (tg-) in this solvent. A comparison is made between the conformation found in Me2SO-d6 and those determined by X-ray diffraction and conformational energy calculations.     

Structure-activity studies with somatostatin: role of lysine in positions 4 and 9 for gastric activity

The gastric inhibitory activity of somatostatin analogues modified in position 4 or 9, was investigated in conscious cats prepared with gastric fistulae. Gastric secretion was stimulated with pentagastrin. Deletion of Lys4, or substitution with an alcoholic (Thr) or acidic (Glu) residue yielded analogues with reduced (10% or less) potency compared with somatostatin. In comparison, [Phe4]somatostatin and modified Phe4 analogues [( p-NH2-Phe4]-, [F5Phe4]- and [Phe4,D-Trp8]somatostatin) were approximately equipotent with somatostatin. The high potency of the Phe4 analogues illustrates that a basic side-chain in position 4 is not essential for gastric activity. In contrast, [Thr9]-, [Glu9]-, [Phe9]- and [p-NH2-Phe9]somatostatin were all inactive (less than 5% potency of somatostatin) in the stomach. Thus the lysyl residue in position 9 is more critical than that in position 4 for somatostatin's gastric activity.     

Characterization of the 31P NMR spectrum of the schistosome vector Biomphalaria glabrata and of the changes following infection by Schistosoma mansoni

The 31P nuclear magnetic resonance (NMR) spectrum of the digestive gland-gonad complex (DGG) of the schistosome vector Biomphalaria glabrata was characterized and the effects of infection by Schistosoma mansoni noted. The in vivo spectrum was comprised of 11 peaks, 5 downfield and 6 upfield of an external 85% phosphoric acid standard. Based on a variety of analytical procedures, the upfield peaks from the standard were demonstrated to be composed of carbamoyl phosphate + a mixture of 3 phosphatides and sphingomyelin, the gamma + beta phosphorus resonances of nucleotide triphosphate (NTP) and nucleotide diphosphate (NDP), respectively, the alpha phosphorus resonances of NTP + NDP, NAD(H) + the phosphorus resonance of uridine phosphate from uridine diphosphoglucose (UDPG), the phosphorus resonance of glucose phosphate from UDPG and, last, the beta phosphorus resonance of NTP. The downfield peaks were assigned as glycerophosphoryl choline, intracellular inorganic phosphate (Pi), sugar phosphates + phosphoryl choline, aminoethyl phosphonate (AEP), and ceramide AEP. T1 values for the in vivo NMR components were determined by inversion recovery. Infection produced distinct alterations in the levels of nonnucleotide components of the in vivo 31P NMR spectrum and the spectra of tissue extracts. Specifically, the levels of phosphonate, phospholipids, and carbamoyl phosphate were markedly reduced, and the relative level of Pi was increased. The potential significance of these changes to the parasite-host relationship was discussed. In contrast, starvation resulted in a decreased level of phosphonate only. The pH of the intact DGG was estimated by titrating the inorganic phosphate component of tissue extracts. The mean pH was 6.9 for both control and infected material.     

The effect of surface curvature on the head-group structure and phase transition properties of phospholipid bilayer vesicles

Proton nuclear magnetic resonance spectra at 360 MHz of small sonicated distearoyl phosphatidylcholine vesicles show easily distinguishable resonances due to choline N-methyl head-group protons located in the inner and outer bilayer halves. A study of the chemical shift of these resonances as a function of temperature reveals that the splitting between them increases below the phase transition. This occurs as a result of an upfield shift of the inner layer resonance at the phase transition. Consideration of the possible causes of this effect results in the conclusion that, at the phase transition, there is a change in the organization of the inner layer head-groups which does not occur for the outer layer head-groups.     

Aromatic residues located close to the active center are essential for the catalytic reaction of flap endonuclease-1 from hyperthermophilic archaeon Pyrococcus horikoshii

Flap endonuclease-1 (FEN-1) possessing 5'-flap endonuclease and 5'-->3' exonuclease activity plays important roles in DNA replication and repair. In this study, the kinetic parameters of mutants at highly conserved aromatic residues, Tyr33, Phe35, Phe79, and Phe278-Phe279, in the vicinity of the catalytic centers of FEN-1 were examined. The substitution of these aromatic residues with alanine led to a large reduction in kcat values, although these mutants retained Km values similar to that of the wild-type enzyme. Notably, the kcat of Y33A and F79A decreased 333-fold and 71-fold, respectively, compared with that of the wild-type enzyme. The aromatic residues Tyr33 and Phe79, and the aromatic cluster Phe278-Phe279 mainly contributed to the recognition of the substrates without the 3' projection of the upstream strand (the nick, 5'-recess-end, single-flap, and pseudo-Y substrates) for the both exo- and endo-activities, but played minor roles in recognizing the substrates with the 3' projection (the double flap substrate and the nick substrate with the 3' projection). The replacement of Tyr33, Phe79, and Phe278-Phe279, with non-charged aromatic residues, but not with aliphatic hydrophobic residues, recovered the kcat values almost fully for the substrates without the 3' projection of the upstream strand, suggesting that the aromatic groups of Tyr33, Phe79, and Phe278-Phe279 might be involved in the catalytic reaction, probably via multiple stacking interactions with nucleotide bases. The stacking interactions of Tyr33 and Phe79 might play important roles in fixing the template strand and the downstream strand, respectively, in close proximity to the active center to achieve the productive transient state leading to the hydrolysis.     

1H NMR (500 MHz) of gene 32 protein--oligonucleotide complexes

In concentrated solutions, gene 32 single-stranded DNA binding protein from bacteriophage T4 (gene 32P) forms oligomers with long rotational correlation times, rendering 1H NMR signals from most of the protons too broad to be detected. Small flexible N- and C-terminal domains are present, however, the protons of which give rise to sharp resonances. If the C-terminal A domain (48 residues) and the N-terminal B domain (21 residues) are removed, the resultant core protein of 232 residues (gene 32P) retains high affinity for ssDNA and remains a monomer in concentrated solution, and most of the proton resonances of the core protein can now be observed. Proton NMR spectra (500 MHz) of gene 32P and its complexes with ApA, d(pA)n (n = 2, 4, 6, 8, and 10), and d(pT)8 show that the resonances of a group of aromatic protons shift upfield upon oligonucleotide binding. Proton difference spectra show that the 1H resonances of at least one Phe, one Trp, and five Tyr residues are involved in the chemical shift changes observed with nucleotide binding. The number of aromatic protons involved and the magnitude of the shifts change with the length of the oligonucleotide until the shifts are only slightly different between the complexes with d(pA)8 and d(pA)10, suggesting that the binding groove accommodates approximately eight nucleotide bases. Many of the aromatic proton NMR shifts observed on oligonucleotide complex formation are similar to those observed for oligonucleotide complex formation with gene 5P of bacteriophage fd, although more aromatic residues are involved in the case of gene 32P.(ABSTRACT TRUNCATED AT 250 WORDS)     

Partial intercalation with DNA of peptides containing two aromatic amino acids

The interactions with DNA of tetrapeptide amides containing lysine at the N-terminal position and aromatic amino acids at the second and fourth positions (Ala at position three), 1-6, have been investigated by nmr, CD, and viscometric methods. Tetrapeptides with N-terminal lysine and a single aromatic amino acid, 7-10, were investigated as controls. Significant decreases in DNA viscosity occurred on addition of 7, with the aromatic group at the second position, but not with any of the other single aromatic amino acid peptides. All of the tetrapeptides with two aromatic groups caused DNA viscosity decreases which were two to three times larger than with 7. Peptides with p-nitrophenylalanine (p-NO2Phe) as the aromatic group were synthesized for nmr studies because of its simpler aromatic nmr spectrum relative to Phe. Large upfield shifts of the aromatic proton signals were obtained when the amino acid in the second position was L-p-NO2Phe, and the fourth position contained either p-NO2Phe or Phe. Such peptides also caused the largest DNA viscosity decreases on complex formation. Smaller upfield shifts of the aromatic signals were obtained when the amino acid in the second position was L-Phe or a D isomer of Phe or p-NO2Phe. With all peptides, larger upfield nmr shifts were obtained with heat-denatured, recooled DNA than with native DNA under the same conditions. As with nmr, CD results are quite different for the peptides with L and D amino acids at the second position. All of the results can be interpreted in terms of a model in which lysine interacts stereospecifically with the backbone in a DNA double helix and the aromatic group at the second position stacks strongly with the base pairs when the amino acid is an L isomer. The aromatic group at the fourth position can also interact with the base pairs, but primarily through a sideways stacking of the aromatic group with base pairs for either L or D isomers. Because of covalent constraints on the separation distance for the two aromatic groups in the tetrapeptides, they must stack on opposite sides of the same base pair in violation of the neighbor exclusion principle observed with classical intercalators. This stacking at the same base pair no doubt accounts for the larger viscosity decreases in DNA with the peptides containing two aromatic groups relative to those with a single aromatic group.(ABSTRACT TRUNCATED AT 400 WORDS)