Search

Yahoo's native advertising (also known as Gemini native) serves billions of ad impressions daily, reaching a yearly run-rate of many hundred of millions USD. Driving the Gemini native models that are used to predict both click probability (pCTR) and conversion probability (pCONV) is offset\ -- a feature enhanced collaborative-filtering (CF) based event prediction algorithm. offset is a one-pass algorithm that updates its model for every new batch of logged data using a stochastic gradient descent (SGD) based approach. Since offset represents its users by their features (i.e., user-less model) due to sparsity issues, rule based hard frequency capping (HFC) is used to control the number of times a certain user views a certain ad. Moreover, related statistics reveal that user ad fatigue results in a dramatic drop in click through rate (CTR). Therefore, to improve click prediction accuracy, we propose a soft frequency capping (SFC) approach, where the frequency feature is incorporated into the offset model as a user-ad feature and its weight vector is learned via logistic regression as part of offset training. Online evaluation of the soft frequency capping algorithm via bucket testing showed a significant $7.3$% revenue lift. Since then, the frequency feature enhanced model has been pushed to production serving all traffic, and is generating a hefty revenue lift for Yahoo Gemini native. We also report related statistics that reveal, among other things, that while users' gender does not affect ad fatigue, the latter seems to increase with users' age.

Cannabis, both herbal and resin, has been the most popular illicit drug in Israel in recent years. Until 2005, the main sources of cannabis resin (known also as hashish) to the Israeli drug market were Lebanon and India. Hashish from these sources can be distinguished by its external appearance. The aim of this study was to determine if there is any difference in the quality of the hashish from each source. For this purpose, we quantified the main cannabinoids, cannabidiol (CBD), Δ9-tetrahydrocannabinol (Δ9-THC), and cannabinol (CBN) of hashish from different police seizures of known origins, Lebanon, India and Morocco, that had been submitted to the Analytical Chemistry Laboratory of the Division of Identification and Forensic Science at the Israeli National Police Headquarters and subsequently to the Hebrew University for quantitative analysis. The results, based on many different seizures, showed that the average CBD content of hashish from Lebanon varied from 5.69% to 12.79% (an average 8.98 ± 0.59%), THC of hashish from Lebanon varied from 0.93% to 4.20% (an average of 2.38 ± 0.27 %), CBD of hashish from Morocco varied from 1.52% to 5.14% (an average of 3.72 ± 0.19%), THC of hashish from Morocco varied from 5.08% to 13.41% (an average of 9.21 ± 0.40%), CBD of hashish from India varied from 0.78% to 13.13% (an average of 4.59 ± 1.07%), and THC of hashish from India varied from 0.53% to 16.45% (an average of 6.35 ± 1.50%), At the same time, several other cannabinoids present in the samples in lower amounts were identified (cannabidivarol, CBDV; cannabicitran; Δ9-tetrahydrocannabivarol, Δ9-THCV; cannabivarol, CBV; cannabicyclol, CBL; cannabichromene, CBC; cannabielsoin, CBE; Δ8-THC; and cannabigerol, CBG). The samples, predominantly from Lebanon, Morocco, and India, were evaluated for chemical phenotype (drug type and fiber type) to determine the geographical origin of these samples.Indian hashish, seized by Czech Republic authorities, was analyzed qualitatively for many cannabinoids and terpenes. In this sample three new cannabinoids were identified: cannabiorcochromene, cis-Δ9-tetrahydrocannabivarol, and cannabinerol.

Finding a maximum cut is a fundamental task in many computational settings, with a central application in wireless networks. Surprisingly, Max-Cut has been insufficiently studied in the classic distributed settings, where vertices communicate by synchronously sending messages to their neighbors according to the underlying graph, known as the \(\mathcal {LOCAL}\) or \(\mathcal {CONGEST}\) models. We amend this by obtaining almost optimal algorithms for Max-Cut on a wide class of graphs in these models. In particular, for any \(\epsilon > 0\), we develop randomized approximation algorithms achieving a ratio of \((1-\varepsilon )\) to the optimum for Max-Cut on bipartite graphs in the \(\mathcal {CONGEST}\) model, and on general graphs in the \(\mathcal {LOCAL}\) model.

Interaction of CC chemokine receptor 5 (CCR5) with the human immunodeficiency virus type 1 (HIV-1) gp120/CD4 complex involves its amino-terminal domain (Nt-CCR5) and requires sulfation of two to four tyrosine residues in Nt-CCR5. The conformation of a 27-residue Nt-CCR5 peptide, sulfated at Y10 and Y14, was studied both in its free form and in a ternary complex with deglycosylated gp120 and a CD4-mimic peptide. NMR experiments revealed a helical conformation at the center of Nt-CCR5(1-27), which is induced upon gp120 binding, as well as a helical propensity for the free peptide. A well-defined structure for the bound peptide was determined for residues 7-23, increasing by 2-fold the length of Nt-CCR5's known structure. Two-dimensional saturation transfer experiments and measurement of relaxation times highlighted Nt-CCR5 residues Y3, V5, P8-T16, E18, I23 and possibly D2 as the main binding determinant. A calculated docking model for Nt-CCR5(1-27) suggests that residues 2-22 of Nt-CCR5 interact with the bases of V3 and C4, while the C-terminal segment of Nt-CCR5(1-27) points toward the target cell membrane, reflecting an Nt-CCR5 orientation that differs by 180° from that of a previous model. A gp120 site that could accommodate (CCR5)Y3 in a sulfated form has been identified. The present model attributes a structural basis for binding interactions to all gp120 residues previously implicated in Nt-CCR5 binding. Moreover, the strong interaction of sulfated (CCR5)Tyr14 with (gp120)Arg440 revealed by the model and the previously found correlation between E322 and R440 mutations shed light on the role of these residues in HIV-1 phenotype conversion, furthering our understanding of CCR5 recognition by HIV-1.

Lod, a city near Tel-Aviv, is considered the main drug distribution center in Israel. A major police undercover operation in Lod, lasting close to a year, was terminated in May 2003. The success or failure of such an operation is frequently measured by the number of arrests made, the hierarchical level of the dealers arrested, the number of drug stations closed down, and the decrease in heroin seizures following the operation. In this work we suggest using an additional parameter, which has a scientific, objective basis, namely, comparing the changes in the average user weight unit ("dose") before and after the operation. We found that prior to the operation the average weight per unit was 1.1 g. Three months after the operation terminated the average weight per unit had decreased to 0.8 g and remained there for at least 4 months before rising again.

Type I interferons (IFNs) are a family of homologous helical cytokines that exhibit pleiotropic effects on a wide variety of cell types, including antiviral activity and antibacterial, antiprozoal, immunomodulatory, and cell growth regulatory functions. Consequently, IFNs are the human proteins most widely used in the treatment of several kinds of cancer, hepatitis C, and multiple sclerosis. All type I IFNs bind to a cell surface receptor consisting of two subunits, IFNAR1 and IFNAR2, associating upon binding of interferon. The structure of the extracellular domain of IFNAR2 (R2-EC) was solved recently. Here we study the complex and the binding interface of IFNalpha2 with R2-EC using multidimensional NMR techniques. NMR shows that IFNalpha2 does not undergo significant structural changes upon binding to its receptor, suggesting a lock-and-key mechanism for binding. Cross saturation experiments were used to determine the receptor binding site upon IFNalpha2. The NMR data and previously published mutagenesis data were used to derive a docking model of the complex with an RMSD of 1 Angstrom, and its well-defined orientation between IFNalpha2 and R2-EC and the structural quality greatly improve upon previously suggested models. The relative ligand-receptor orientation is believed to be important for interferon signaling and possibly one of the parameters that distinguish the different IFN I subtypes. This structural information provides important insight into interferon signaling processes and may allow improvement in the development of therapeutically used IFNs and IFN-like molecules.

The potent antiviral and antiproliferative activities of human type I interferons (IFNs) are mediated by a single receptor comprising two subunits, IFNAR1 and IFNAR2. The structure of the IFNAR2 IFN binding ectodomain (IFNAR2-EC), the first helical cytokine receptor structure determined in solution, reveals the molecular basis for IFN binding. The atypical perpendicular orientation of its two fibronectin domains explains the lack of C domain involvement in ligand binding. A model of the IFNAR2-EC/IFNα2 complex based on double mutant cycle-derived constraints uncovers an extensive and predominantly aliphatic hydrophobic patch on the receptor that interacts with a matching hydrophobic surface of IFNα2. An adjacent motif of alternating charged side chains guides the two proteins into a tight complex. The binding interface may account for crossreactivity and ligand specificity of the receptor. This molecular description of IFN binding should be invaluable for study and design of IFN-based biomedical agents.

The Fv is the smallest antigen binding fragment of the antibody and is made of the variable domains of the light and heavy chains, V L and V H , respectively. The 26-kDa Fv is amenable for structure determination in solution using multi-dimensional hetero-nuclear NMR spectroscopy. The human monoclonal antibody 447-52D neutralizes a broad spectrum of HIV-1 isolates. This anti-HIV-1 antibody elicited in an infected patient is directed against the third variable loop (V3) of the envelope glycoprotein (gp120) of the virus. The V3 loop is an immunodominant neutralizing epitope of HIV-1. To obtain the 447-52D Fv for NMR studies, an Escherichia coli bicistronic expression vector for the heterodimeric 447-52D Fv and vectors for single chain Fv and individually expressed V H and V L were constructed. A pelB signal peptide was linked to the antibody genes to enable secretion of the expressed polypeptides into the periplasm. For easy cloning of any antibody gene without potential modification of the antibody sequence, restriction sites were introduced in the pelB sequence and following the termination codon. A set of oligonucleotides that prime the leader peptide genes of all potential antibody human antibodies were designed as backward primers. The forward primers for the V L and V H were based on constant region sequences. The 447-52D Fv could not be expressed either by a bicistronic vector or as single chain Fv, probably due to its toxicity to Escherichia coli . High level of expression was obtained by individual expression of the V H and the V L chains, which were then purified and recombined to generate a soluble and active 447-52D Fv fragment. The V L of mAb 447-52D was uniformly labeled with 13 C and 15 N nuclei (U– 13 C/ 15 N). Preliminary NMR spectra demonstrate that structure determination of the recombinant 447-52D Fv and its complex with V3 peptides is feasible.

The human interferon receptor (IFNAR) mediates the antiviral and antiproliferative activities of type I interferons (IFNs). This receptor is comprised of subunits IFNAR1 and IFNAR2, the latter exhibiting nanomolar affinity for IFNs. Here the extracellular domain of IFNAR2 (IFNAR2-EC), a soluble 25 kDa IFN-binding polypeptide, and its complex with IFN-alpha 2 were studied using multidimensional NMR. IFNAR2-EC is comprised of two fibronectin-III (FN-III) domains connected by a helical hinge region. The deduced global fold was utilized to improve the alignment of IFNAR2-EC against structurally related receptors and to model its structure. A striking feature of IFNAR2-EC is the limited and localized deviations in chemical shifts exhibited upon ligand binding, observed for only 15% of its backbone (1)H and (15)N nuclei. Analysis of these deviations maps the IFN-alpha 2 binding site upon IFNAR2-EC to a contiguous surface on the N-terminal domain, including the S3-S4 loop (residues 44-53), the S5-S6 loop and S6 beta-strand (residues 74-82), and the S7 beta-strand and the hinge region (residues 95-105). The C-terminal domain contributes only marginally to ligand binding, and no change in the hypothesized interdomain interface is observed. The proposed binding domain encompasses all residues implicated by mutagenesis studies in IFN binding, and suggests adjacent residues cooperate in forming the binding surface. D(2)O-exchange experiments indicate that binding of IFN-alpha2 induces tightening of the N-terminal domain of IFNAR2-EC. This increase in receptor rigidity may play an important role in initiating the intracellular stage of the IFN signaling cascade.

A cis proline turn linking two β-hairpin strands in the solution structure of an antibody-bound HIV-1 IIIB V3 peptide

The construction, expression, and purification of an active Fv fragment of the 0.5beta monoclonal human immunodeficiency virus type 1 (HIV-1) neutralizing antibody is reported. The interaction between the Fv fragment and the RP135 peptide derived from the V3 loop of gp120 from HIV-1IIIB was studied by varying the salt concentration and by mutating arginine residues in the peptide. The mutations R4A, R8A and R11A (which correspond to residues 311, 315, and 318 in gp120 of HIV-1IIIB) reduce the binding free energy by 0.22 (+/- 0. 20), 4.32 (+/- 0.16), and 1.58 (+/- 0.17) kcal mol-1, respectively. The salt-dependent components of their contributions to binding are 0.02 (+/- 0.22), -0.55 (+/- 0.18), and -0.97 (+/- 0.19) kcal mol-1, respectively. The magnitudes of the mutational effects and the extent of shielding by 1 M NaCl suggest that Arg-8 is involved in a buried salt bridge in the peptide-Fv fragment complex, whereas Arg-11 is involved in a more solvent-exposed electrostatic interaction.

The 24-amino-acid peptide RP135 (NNTRKSIRIQRGPGRAFVTIGKIG) corresponds in its amino acid sequence to the principal neutralizing determinant of the human immunodeficiency virus type-1, IIIB isolate (HIV-1IIIB, residues 308-331 of the envelope glycoprotein gp120). In order to map the antigenic determinant recognized by 0.5 beta, the complex of RP135 with an anti-gp120 HIV neutralizing antibody, 0.5 beta, which cross reacts with the peptide, was studied by using two-dimensional NMR spectroscopy. A combination of homonuclear Hartmann Hahn two-dimensional experiment and roating-frame Overhauser enhancement spectroscopy of the Fab/peptide complex measured in H2O was used to eliminate the resonances of the Fab and the tightly bound peptide residues and to obtain sequential assignments for those parts of the peptide which retain considerable mobility upon binding. In this manner, a total of 14 residues (Ser6-Thr19) were shown to be part of the antigenic determinant recognized by the antibody 0.5 beta. Lys5 and Ile20 were found to retain considerable mobility in the bound peptide while their amide protons undergo significant change in chemical shift upon binding. This observation suggests that these two residues are at the boundaries of the determinant recognized by the antibody. Competitive binding experiments using truncated peptides strongly support the NMR observations.

Type I interferons (IFNs) make up a family of homologous helical cytokines initiating strong antiviral and antiproliferative activity. All type I IFNs bind to a common cell surface receptor consisting of two subunits, IFNAR1 and IFNAR2, associating upon binding of interferon. We studied intermolecular interactions between IFNAR2-EC and IFNalpha2 using asymmetric reverse-protonation of the different complex components and two-dimensional homonuclear NOESY. This new approach revealed with an excellent signal-to-noise ratio 24 new intermolecular NOEs between the two molecules despite the low concentration of the complex (0.25 mM) and its high molecular mass (44 kDa). Sequential and side chain assignment of IFNAR2-EC and IFNalpha2 in their binary complex helped assign the intermolecular NOEs to the corresponding protons. A docking model of the IFNAR2-EC-IFNalpha2 complex was calculated on the basis of the intermolecular interactions found in this study as well as four double mutant cycle constraints, previously observed NOEs between a single pair of residues and the NMR mapping of the binding sites on IFNAR2-EC and IFNalpha2. Our docking model doubles the buried surface area of the previous model and significantly increases the number of intermolecular hydrogen bonds, salt bridges, and van der Waals interactions. Furthermore, our model reveals the participation of several new regions in the binding site such as the N-terminus and A helix of IFNalpha2 and the C domain of IFNAR2-EC. As a result of these additions, the orientation of IFNAR2-EC relative to IFNalpha2 has changed by 30 degrees in comparison with a previously calculated model that was based on NMR mapping of the binding sites and double mutant cycle constraints. In addition, the new model strongly supports the recently proposed allosteric changes in IFNalpha2 upon binding of IFNAR1-EC to the binary IFNalpha2-IFNAR2-EC complex.

3,4-Methylenedioxymethamphetamine (MDMA) tablets known as "ecstasy" became a very popular drug amongst Israeli youth in the last decade. The ecstasy tablets have a simple design impressed on them (logos) making it relatively easy to distinguish between various logos. The life expectancy of ecstasy tablet logos, defined as the period between the first seizure by the police of a certain logo until the last seizure of the same logo, was monitored during the years 2001-2003. During this time interval, 58 different tablet logos were seized. A total of 26 logos, defined as common logos with at least 10 independent seizures, were observed. At any given time interval during this period, 8-10 common logos were found with an average life expectancy of approximately 9 months. Five of the observed 26 common logos were defined as the most common logos that appeared in at least 200 independent seizures each. Plots of the number of seizures and number of tablets seized as a function of time are presented and discussed as well as explanations for the high turnover rate of any given logo.

A heroin distribution chain was simulated by taking three different seizures and preparing four additional samples from each seizure by adding a paracetamol-caffeine mixture in varying amounts, resulting in three different batches each composed of five samples. All of the samples from the three batches were analyzed using HPLC with a UV-PDA detector at a wavelength of 230 nm. The area ratio of various opium alkaloids, acetylation products and components were compared. From the results of the UV area ratios, the fifteen samples could readily be separated into three batches of five samples, with each batch of five samples having a common origin.

The potent antiviral and antiproliferative activities of human type I interferons (IFNs) are mediated by a single receptor comprising two subunits, IFNAR1 and IFNAR2. The structure of the IFNAR2 IFN binding ectodomain (IFNAR2-EC), the first helical cytokine receptor structure determined in solution, reveals the molecular basis for IFN binding. The atypical perpendicular orientation of its two fibronectin domains explains the lack of C domain involvement in ligand binding. A model of the IFNAR2-EC/IFNalpha2 complex based on double mutant cycle-derived constraints uncovers an extensive and predominantly aliphatic hydrophobic patch on the receptor that interacts with a matching hydrophobic surface of IFNalpha2. An adjacent motif of alternating charged side chains guides the two proteins into a tight complex. The binding interface may account for crossreactivity and ligand specificity of the receptor. This molecular description of IFN binding should be invaluable for study and design of IFN-based biomedical agents.

Most human immunodeficiency virus type 1 (HIV-1) neutralizing antibodies in infected individuals and in immunized animals are directed against the third variable loop (V3) of the envelope glycoprotein (gp120) of the virus. This loop plays a crucial role in phenotypic determination, cytopathicity (syncytium induction), and coreceptor usage of HIV-1. The human monoclonal antibody 447-52D was found to neutralize a broad spectrum of HIV-1 strains. In order to solve the solution structure of the V3MN peptide bound to the 447-52D Fab fragment by NMR, large quantities of labeled peptide and a protocol for the purification of the Fab fragment were needed. An expression plasmid coding for the 23-residue V3 peptide of the HIV-1MN strain (V3MN peptide, YNKRKRIHIGPGRAFYTTKNIIG) linked to a derivative of the RNA-binding domain of hnRNCP1 was constructed. The fusion protein attached to the V3 peptide prevents its degradation. Using this system, U-15N, U-13C,15N, and U-13C,15N, 50% 2H labeled fusion protein molecules were expressed in Escherichia coli grown on rich Celtone medium with yields of about 240 mg/liter. The V3MN peptide was released by CNBr cleavage and purified by RP-HPLC, giving final yields of 6–13 mg/liter. This expression system is generally applicable for biosynthesis of V3-related peptides and was also used to prepare the V3JR-FL. The 447-52D Fab fragment was obtained by a short enzymatic papain cleavage of the whole antibody. Preliminary NMR spectra demonstrate that full structural analysis of the V3MN complexed to the 447-52D Fab is feasible. This system enables studies of the same epitope bound to different HIV-1 neutralizing antibodies.

To increase our understanding of the molecular basis for antibody specificity and for the cross-reactivity of antipeptide antibodies with native proteins, it is important to study the three-dimensional structure of antibody complexes with their peptide antigens. For this purpose it may not be necessary to solve the structure of the whole antibody complex but rather to concentrate on elucidating the combining site structure, the interactions of the antibody with its antigen, and the bound peptide conformation. To extract the information about antibody–peptide interactions and intramolecular interactions in the bound ligand from the complicated and unresolved spectrum of the Fab–peptide complex (Fab: antibody fragment made of Fv—the antibody fragment composed of the variable regions of the light and heavy chains forming a single combining site for the antigen—the light chain, and the first heavy chain constant regions), an nmr methodology based on measurements of two-dimensional transferred nuclear Overhauser effect (NOE) difference spectra was developed. Using this methodology the interactions of three monoclonal antibodies with a cholera toxin peptide were studied. The observed interactions were assigned to the antibody protons involved by specific deuteration of aromatic amino acids and specific chain labeling, and by using a predicted model for the structure of the antibody combining site. The assigned NOE interactions were translated to restraints on interproton distances in the complex that were used to dock the peptide into calculated models for the antibodies combining sites. Comparison of the interactions of three antibodies against a cholera toxin peptide (CTP3). which differ in their cross-reactivity with the toxin, yields information about the size and conformation of antigenic determinants recognized by the antibodies, the structure of their combining sites, and relationships between antibodies' primary structure and their interactions with peptide antigens. © 1994 John Wiley & Sons, Inc.

To increase our understanding of the molecular basis for antibody specificity and for the cross-reactivity of anti-peptide antibodies with native proteins it is important to study the three-dimensional structure of antibody complexes with their peptide antigens. For this purpose it may not be necessary to solve the structure of the whole antibody complex but rather to concentrate on elucidating the combining site structure, the interactions of the antibody with its antigen and the bound peptide conformation. We have developed an NMR methodology based on two-dimensional difference spectrum measurements which extract the information concerning antibody-peptide interactions and intramolecular interactions in the bound ligand from the crowded and unresolved spectrum of the Fab complex. These measurements yield restraints on interproton distances in the complex which are used to dock the peptide into calculated models for the antibodies' combining sites. Comparison of the interactions of three antibodies against a cholera toxin peptide (CTP3), which differ in their cross-reactivity with the toxin, yields information about the size and conformation of antigenic determinants recognized by antibodies, the structure of their combining sites and relationships between antibodies' primary structure, and their interactions with peptide antigens.

During 1992, a survey was performed on heroin powder exhibits seized throughout Israel. The main objectives were to detect salt forms and additives and compare the powders for intelligence and investigation purposes. A total of 3817 exhibits were scanned by FTIR; five major groups were distinguished: heroin free base, heroin base with caffeine, heroin base with paracetamol (acetaminophen), heroin base containing both caffeine and paracetamol, and a mixture of heroin base, heroin HCl and caffeine. In addition, 419 exhibits collected during a two month period were compared amongst themselves and categorized into sub-groups based on the similarity of their spectra. Based on this data, it was possible to define 116 different sub-groups, of which 56 sub-groups contained from 2 to 47 samples. Of the total number of cases tested, 78% were in the weight range 1 mg-1 g, that is, user quantities. Results were tabulated on both a geographical and monthly basis. Differences were found in the distribution of additives between different geographical areas of the country.

The interactions between the aromatic amino acids of two monoclonal antibodies (TE32 and TE33) with specific amino acid residues of a peptide of cholera toxin (CTP3) have been determined by two-dimensional (2D) transferred NOE difference spectroscopy. Aromatic amino acids are found to play an important role in peptide binding. In both antibodies two tryptophan and two tyrosine residues and one histidine residue interact with the peptide. In TE33 there is an additional phenylalanine residue that also interacts with the peptide. The residues of the CTP3 peptide that have been found to interact with the antibody are val 3, pro 4, gly 5, gln 7, his 8, and asp 10. We have determined the amino acid sequences of the two antibodies by direct mRNA sequencing. Computerized molecular modeling has been used to build detailed all-atom models of both antibodies from the known conformations of other antibodies. These models allow unambiguous assignment of most of the antibody residues that interact with the peptide. A comparison of the amino acid sequences of the two anti-CTP3 antibodies with other antibodies from the same gene family reveals that the majority of the aromatic residues involved in the binding of CTP3 are conserved although these antibodies have different specificities. This similarity suggests that these aromatic residues create a general hydrophobic pocket and that other residues in the complementarity-determining regions (CDRs) modulate the shape and the polarity of the combining site to fit the specific antigens.

The exocytosis of catecholamines by chromaffin cells following stimulation (e.g. by acetylcholine) is accompanied by a rise in the level of intracellular free Ca2+. Actually, secretion can be induced merely by making the cells leaky to Ca2+ from the external medium. We have recently demonstrated that secretion can be increased by the introduction of DNase-I, the F-actin depolymerizing agent, or of heavy meromyosin, the enzymatically active fragment of myosin. Suspecting that these changes might be associated with a higher intracellular level of Ca2+, we now have measured the influx of 45Ca2+ into chromaffin cells which have undergone fusion with DNase-I- or with heavy meromyosin-loaded liposomes. In both cases, a marked increase in Ca2+ uptake has been observed, which could be abolished by Co2+ ions (a Ca2+ channel blocker), suggesting an intimate involvement of the cellular actomyosin system in the process of Ca2+ ions transport through the Ca2+ channels of the plasma membrane.

The interactions between a peptide of cholera toxin and the aromatic amino acids of the TE33 antipeptide antibody, cross-reactive with the toxin, have been studied by NOESY difference spectroscopy. The 2D difference between the NOESY spectrum of the Fab with a 4-fold excess of the peptide and that of the peptide-saturated Fab reveals cross-peaks growing with excess of the peptide. These cross-peaks are due to magnetization transfer between the Fab and neighboring bound peptide protons, and a further transfer to the free peptide protons by exchange between bound and free peptide (transferred NOE). Additional cross-peaks appearing in the difference spectrum are due to a combination of intramolecular interactions between bound peptide protons and exchange between bound and free peptide. Assignment of cross-peaks is attained by specific deuteration of antibody aromatic amino acids using also the resonance assignment of the free peptide, deduced from the COSY spectrum of the peptide solution. The antibody combining site is found to be highly aromatic. We have identified one or two histidine, two tyrosine, and two tryptophan residues and one phenylalanine residue of the antibody interacting with valine-3, proline-4, glycine-5, glutamine-7, histidine-8, and aspartate-10 of the peptide. The 2D TRNOE difference spectroscopy can be used to study protein-ligand interactions, given that the ligand off rate is fast relative to the spin-lattice relaxation time of the protein and ligand protons (about 1 s). The resolution obtained in the difference spectra implies that the technique is equally applicable for studying proteins having a molecular weight larger than 50,000.(ABSTRACT TRUNCATED AT 250 WORDS)