New crystal structures of ColE1 Rom and variants resulting from mutation of a surface exposed residue: Implications for RNA-recognition.

In ColE1, the plasmid encoded RNA one modulator (Rom) protein, which is also referred to as Rop, specifically binds and stabilizes an intermediate RNA loop-loop kissing structure formed between the plasmid encoded transcripts RNA I and RNA II and thereby acts as an auxiliary repressor of replication. Rom folds into a homodimeric, cylindrically packed four helix bundle with an exact twofold symmetry axis (Banner et al., J Mol Biol 1987;196:657-675; Eberle et al., J Biomol 1991;1:71-83). Previous studies (Castagnoli et al., EMBO J 1989;8:621-629; Predki et al., Cell 1995;80:41-50) have localized the RNA binding surface to the H1/H1' face of the helical bundle and found Phe14 to be a key determinant of the binding affinity and specificity for RNA kissing complexes. To investigate the role of Phe14 in RNA recognition, we have determined high-resolution crystal structures of two point mutants of Rom (F14Y and F14W), as well as a high-resolution structure of a crystal form of Rom in which the dimer comprises the asymmetric unit. Although the structures of F14Y and F14W share a very high degree of structural identity with that of the wild-type protein and each other, differences are observed between the three polypeptide chains found in the asymmetric unit of each crystal in the packing of the tryptophan and tyrosine side chains at position 14, as well as some of the other surface exposed side chains of key amino acids involved in RNA binding. In both the wild-type Rom and mutant structures, crystal packing forces can break the exact twofold symmetry of the dimer and influence the conformation of the side chains presented on the H1/H1' face of Rom. Since the new structures show such a high degree of structural identity, the disruption in RNA binding observed for the mutant proteins can be attributed specifically to the chemical nature of the side chain at position 14. Moreover, the fact that even subtle changes in the side chain at position 14 cannot be compensated for by the apparent flexibility of this side chain suggests a highly constrained packing of this residue in the RNA-protein complex.

NMR analysis of rhodopsin-transducin interactions.

Heterotrimeric G-protein activation by an agonist-stimulated G-protein coupled receptor (R*) requires the propagation of structural signals from the receptor interacting surfaces to the guanine nucleotide-binding pocket. Employing high-resolution NMR methods, we are probing heterotrimer-associated and rhodopsin-stimulated changes in an isotope-labeled G-protein alpha-subunit (G(alpha)). A key aspect of the work involves the trapping and interrogation of discrete R*-bound conformations of G(alpha). Our results demonstrate that functionally important changes in G(alpha) structure and dynamics can be detected and characterized by NMR, enabling the generation of robust models for the global and local structural changes accompanying signal transfer from R* to the G-protein.

Evidence for structural changes in carboxyl-terminal peptides of transducin alpha-subunit upon binding a soluble mimic of light-activated rhodopsin.

Although a high-resolution crystal structure for the ground state of rhodopsin is now available, portions of the cytoplasmic surface are not well resolved, and the structural basis for the interaction of the cytoplasmic loops with the retinal G-protein transducin (G(t)) is still unknown. Previous efforts aimed at the design, construction, and functional characterization of soluble mimics for the light-activated state of rhodopsin have shown that grafting defined segments from the cytoplasmic region of bovine opsin onto a surface loop in a mutant form of thioredoxin (HPTRX) is sufficient to confer partial G(t) activating potential [Abdulaev et al. (2000) J. Biol. Chem. 275, 39354-39363]. To assess whether these designed mimics could provide a structural insight into the interaction between light-activated rhodopsin and G(t), the ability of an HPTRX fusion protein comprised of the second (CD) and third (EF) cytoplasmic loops (HPTRX/CDEF) to bind G(t) alpha-subunit (G(t)(alpha)) peptides was examined using nuclear magnetic resonance (NMR) spectroscopy. Transfer NOESY (TrNOESY) experiments show that an 11 amino acid peptide corresponding to the carboxyl terminus of G(t)(alpha) (GtP), as well as a "high-affinity" peptide analogue, HAP1, binds to HPTRX/CDEF in the fast-exchange regime and undergoes similar, subtle structural changes at the extreme carboxyl terminus. Observed TrNOEs suggest that both peptides when bound to HPTRX/CDEF adopt a reverse turn that is consistent with the C-cap structure that has been previously reported for the interaction of GtP with the light-activated signaling state, metarhodopsin II (MII). In contrast, TrNOESY spectra provide no evidence for structuring of the amino terminus of either GtP or HAP1 when bound to HPTRX/CDEF, nor do the spectra show any measurable changes in the CD and EF loop resonances of HPTRX/CDEF, which are conformationally dynamic and significantly exchange broadened. Taken together, the NMR observations indicate that HPTRX/CDEF, previously identified as a functional mimic of MII, is also an approximate structural mimic for this light-activated state of rhodopsin.

S(3)E-E.COSY methods for the measurement of (19)F associated scalar and dipolar coupling constants.

A (1)H-(19)F spin state selective excitation (S(3)E) pulse sequence element has been applied in combination with (1)H homonuclear mixing to create E.COSY-type experiments designed to measure scalar J(HF2') and J(HH2') and residual dipolar D(HF2') and D(HH2') couplings in 2'-deoxy-2'-fluoro-sugars. The (1)H-(19)F S(3)E pulse sequence element, which resembles a simple INEPT sequence, achieves spin-state-selective correlation between geminal (1)H-(19)F spin pairs by linear combination of in-phase (19)F magnetization and anti-phase magnetization evolved from (1)H. Since the S(3)E sequence converts both (19)F and (1)H steady-state polarization into observable coherences, an approximately twofold signal increase is observed for fully relaxed (1)H-(19)F spin pairs with respect to a standard (1)H coupled (19)F 1D experiment. The improved sensitivity and resolution afforded by the use of (1)H-(19)F S(3)E E.COSY-type experiments for measuring couplings is demonstrated on the nucleoside 9-(2',3'-dideoxy-2'-fluoro-beta-D-threo-pentofuranosyl)adenine (beta-FddA) and on a selectively 2'-fluorine labeled 21mer RNA oligonucleotide.

Steady-state kinetic characterization of substrates and metal-ion specificities of the full-length and N-terminally truncated recombinant human methionine aminopeptidases (type 2).

The steady-state kinetics of a full-length and truncated form of the type 2 human methionine aminopeptidase (hMetAP2) were analyzed by continuous monitoring of the amide bond cleavage of various peptide substrates and methionyl analogues of 7-amido-4-methylcoumarin (AMC) and p-nitroaniline (pNA), utilizing new fluorescence-based and absorbance-based assay substrates and a novel coupled-enzyme assay method. The most efficient substrates for hMetAP2 appeared to be peptides of three or more amino acids for which the values of k(cat)/K(m) were approximately 5 x 10(5) M(-1) min(-1). It was found that while the nature of the P1' residue of peptide substrates dictates the substrate specificity in the active site of hMetAP2, the P2' residue appears to play a key role in the kinetics of peptidolysis. The catalytic efficiency of dipeptide substrates was found to be at least 250-fold lower than those of the tripeptides. This substantially diminished catalytic efficiency of hMetAP2 observed with the alternative substrates MetAMC and MetpNA is almost entirely due to the reduction in the turnover rate (k(cat)), suggesting that cleavage of the amide bond is at least partially rate-limiting. The 107 N-terminal residues of hMetAP2 were not required for either the peptidolytic activity of the enzyme or its stability. Steady-state kinetic comparison and thermodynamic analyses of an N-terminally truncated form and full-length enzyme yielded essentially identical kinetic behavior and physical properties. Addition of exogenous Co(II) cation was found to significantly activate the full-length hMetAP2, while Zn(II) cation, on the other hand, was unable to activate hMetAP2 under any concentration that was tested.

Measurement and application of 1H-19F dipolar couplings in the structure determination of 2'-fluorolabeled RNA.

Residual dipolar couplings can provide powerful restraints for determination and refinement of the solution structure of macromolecules. The application of these couplings in nucleic acid structure elucidation can have an especially dramatic impact, since they provide long-range restraints, typically absent in NOE and J-coupling measurements. Here we describe sensitive X-filtered-E.COSY-type methods designed to measure both the sign and magnitude of long-range 1H-19F dipolar couplings in selectively fluorine labeled RNA oligonucleotides oriented in solution by a liquid crystalline medium. The techniques for measuring 1H-19F dipolar couplings are demonstrated on a 21-mer RNA hairpin, which has been specifically labeled with fluorine at the 2'-hydroxyl position of three ribose sugars. Experimentally measured 1H-19F dipolar couplings for the 2'-deoxy-2'-fluoro-sugars located in the helical region of the RNA hairpin were found to be in excellent agreement with values predicted using canonical A-form helical geometry, demonstrating that these couplings can provide accurate restraints for the refinement of RNA structures determined by NMR.

Sulfoxide-controlled S(N)2' displacements between cyanocuprates and epoxy vinyl sulfoxides.

Two short and convergent routes have been devised for the preparation of enantiomerically pure acyclic epoxy vinyl sulfoxides. These substrates undergo highly regio- and stereoselective S(N)2' displacements with lithium cyanocuprates to give alpha'-alkylated, gamma-oxygenated Z alpha,beta-unsaturated sulfoxides in moderate to good yields and with good to excellent diastereoselectivities. The absolute configuration of the newly formed carbon-carbon bond is primarily controlled by the chiral sulfur atom, which in a nonreinforcing situation can override the intrinsic anti tendency of the vinyl oxirane moiety and forces the cuprate to undergo syn addition. The hydroxy vinyl sulfoxide functionality of the resulting adducts should allow for subsequent asymmetric transformations thus enhancing the synthetic usefulness of this methodology.

Mechanism of neomycin and Rev peptide binding to the Rev responsive element of HIV-1 as determined by fluorescence and NMR spectroscopy.

Rev is an essential HIV-1 regulatory protein that binds the Rev responsive element (RRE) within the env gene of the HIV-1 RNA genome and is involved in transport of unspliced or partially spliced viral mRNA from the cell nucleus to the cytoplasm. Previous studies have shown that a short alpha-helical peptide derived from Rev (Rev 34-50), and a truncated form of the RRE sequence provide a useful in vitro system to study this interaction while still preserving the essential aspects of the native complex. We have selectively incorporated the fluorescent probe 2-aminopurine 2'-O-methylriboside (2-AP) into the RRE sequence in nonperturbing positions (A68 and U72) such that the binding of both Rev peptide and aminoglycoside ligands could be characterized directly by fluorescence methods. Rev peptide binding to the RRE-72AP variant resulted in a 2-fold fluorescence increase that provided a useful signal to monitor this binding interaction (K(D) = 20 +/- 7 nM). Using stopped-flow kinetic measurements, we have shown that specific Rev peptide binding occurs by a two-step process involving diffusion-controlled encounter, followed by isomerization of the RNA. Using the RRE-68AP and -72AP constructs, three classes of binding sites for the aminoglycoside neomycin were unambiguously detected. The first site is noninhibitory to Rev binding (K(D) = 0.24 +/- 0.040 microM), the second site inhibited Rev binding in a competitive fashion (K(D) = 1. 8 +/- 0.8 microM), and the third much weaker site (or sites) is attributed to nonspecific binding (K(D) >/= 40 microM). Complementary NMR measurements have shown that neomycin forms both a specific binary complex with RRE and a specific ternary complex with RRE and Rev. NMR data further suggest that neomycin occupies a similar high-affinity binding site in both the binary and ternary complexes, and that this site is located in the lower stem region of RRE.

[Staphylococcus epidermidis infective endocarditis after mitral surgery, successfully treated with aspirin and antibiotics: a case report]. / Endocardite infectieuse à Staphylococcus epidermidis après plastie mitrale, traitée avec succès par aspirine et antibiotiques: à propos d'un cas.

We report a case of infectious endocarditis from Staphylococcus epidermidis that occurred early after mitral valve repair (one month), suggested by fever of 38.5 degrees C and valvular vegetations of less than 10 mm on the mitral valve. In the absence of standard recommendations (medical or surgical) in treating patients with infectious endocarditis occurring after mitral valve repair, and in the absence of complications, a medical regimen was chosen associating aspirin (anti-aggregant dosages) with antibiotics. The benefit of aspirin in endocarditis has been demonstrated in experimental studies with regards to valvular vegetations and embolic risk but remains to be studied in human clinical trials. With the association of aspirin (100 mg/d) and triple antibiotic therapy (rifampicin 1200 mg/d, vancomycin 2 g/d, gentamycin 180 mg/d), the clinical status improved with complete regression of vegetations in less than 24 days and the absence of recurrence at one-year follow-up.

Evaluation of compensation filters in pedal radiographs.

Compensation filters allow increased visibility of detail in chest, shoulder, spine, hip, knee, and foot radiographs. This study examines use of an anatomic compensation filter to improve imaging in pedal radiographs. Anteroposterior radiographs were obtained of 25 cadaveric feet at two settings with and without the compensation filter. Densitometer readings were taken at ten forefoot anatomic sites. The compensation filter produced statistically significant reductions in densitometer readings at all anatomic sites and at both radiographic settings. Filtration improved imagery of bony structures, provided excellent soft-tissue visualization, and lowered patient exposure.

[Cavo-pulmonary anastomosis associated with intracardiac repair of Ebstein anomaly. Value in high-risk patients]. / Anastomose cavo-bipulmonaire associée à la réparation intracardiaque de la maladie d'Ebstein. Intérêt chez les patients à haut risque.

The surgical prognosis of Ebstein's anomaly depends on the quality of tricuspid valve repair and right ventricular function. In patients with right ventricular failure, a decrease in afterload was attained by a cavo-bipulmonary anastomosis associated with the intraventricular repair. Fifty-nine out of 111 patients operated for Ebstein's anomaly were considered to be at high risk and were selected for this study. The inclusion criteria were one or more of the following factors: massive tricuspid regurgitation, extensive atrialisation of the right ventricle, poor right ventricular function, chronic atrial fibrillation. This population was divided into two groups with the same preoperative features: Group I (45 operated patients: tricuspid valvuloplasty with longitudinal plicature of the right ventricule: Group II (14 operated patients): same intracardiac repair as Group I and associated cavo-bipulmonary anastomosis. The operative mortality was 24% (11/45) in Group I and 7% (1/14) in Group II (p < 0.05). The 5 year actuarial was 68.6% in Group I and 61.8% in Group II (NS). The reoperation rate was 11% (5/45) in Group I and 0% in Group II. In Group II, the persistence of significant tricuspid regurgitation was better tolerated and the frequence of reoperation was decreased with respect to Group I. The authors conclude that high risk patients with Ebstein's anomaly have a lower operative mortality and improved functional tolerance when there is persistent tricuspid regurgitation after cavo-bipulmonary anastomosis.

Bi-directional cavopulmonary shunt associated with ventriculo and valvuloplasty in Ebstein's anomaly: benefits in high risk patients.

OBJECTIVE: The prognosis for surgical repair of Ebstein's anomaly depends on the tricuspid valve repair and on the right ventricular function. In order to decrease the preload of the compromised right ventricle, a bi-directional cavopulmonary shunt was added to the intracardiac repair. METHODS: Among 113 patients operated on for Ebstein' s anomaly at our institution from 1980-1997, a cohort of 60 patients was selected for high risk for postoperative complications. Patients presented with one or more several criteria: massive tricuspid valve dysfunction, extended atrialized right ventricle, poor right ventricular contractility, or long standing atrial fibrillation. After prior informed consent, this cohort was divided into two groups. Both groups had similar preoperative clinical patterns: Group I (45 patients), surgical technique included longitudinal right ventricular plication and tricuspid valve valvuloplasty; Group II (15 patients), where the surgical technique was similar to Group I except a bi-directional cavopulmonary shunt was added at the end of the procedure. RESULTS: Operative mortality was 24% (11/45) in Group I and 0% (0/15) in Group II (P < 0.05). The survival at 5 years was 66.1 +/- 14% in Group I and 80 +/- 16% in Group II (not significant). Reoperation rate was 11% (5/45) in Group I and 0% (0/15) in Group II. No deleterious effects of the bi-directional cavopulmonary shunts were observed clinically. Residual tricuspid valve insufficiency rate was 26% in both groups. However, patients with the bi-directional cavopulmonary shunt had a better tolerance and have not needed reoperations to date. CONCLUSIONS: In high risk patients with Ebstein's anomaly, an associated bi-directional cavopulmonary shunt seems to offer several distinct advantages including decreased operative mortality and better tolerance of the residual tricuspid valve dysfunction.

A New Experiment for the Measurement of nJ(C,P) Coupling Constants Including 3J(C4'i,Pi) and 3J(C4'i,Pi+1) in Oligonucleotides.

A new experiment for the measurement of nJ(C,P) coupling constants along the phosphodiester backbone in RNA and DNA based on a quantitative-J HCP experiment is presented. In addition to coupling constants, in which a carbon atom couples to only one phosphorus atom, both the intraresidual 3J(C4'i,Pi) and the sequential 3J(C4'i,Pi+1) for the C4' resonances that couple to two phosphorus atoms can be obtained. Coupling constants obtained by this new method are compared to values obtained from the P-FIDS experiment. Together with 3J(H,P) coupling constants measured using the P-FIDS experiment, the backbone angles ß and ∈ can be determined.

Dynamic cardiomyoplasty: clinical follow-up at 12 years.

OBJECTIVE: The purpose of this study is to evaluate the long-term outcome of dynamic cardiomyoplasty. This surgical technique was conceived to assist the failing heart. The many proposed mechanisms of action of cardiomyoplasty are: (1) systolic assist; (2) limitation of ventricular dilation; (3) reduction of ventricular wall stress (sparing effect); (4) ventricular remodeling with an active girdling effect; (5) angiogenesis; and (6) a neurohumoral effect. METHODS: We investigated 95 patients in our hospital undergoing this procedure due to severe chronic heart failure, refractory to optimal medical treatment. Patients had a mean age of 51 +/- 12 years. The etiology of heart failure was ischemic 55%, idiopathic 34%, ventricular tumor 6%, and other 5%. The mean follow-up was 44 months. RESULTS: The mean New York Heart Association (NYHA) functional class improved postoperatively from 3.2 to 1.8. Average radioisotopic left ventricular (LV) ejection fraction increased from 17 +/- 5 to 27 +/- 4% (P < 0.05). Stroke volume index increased from 32 +/- 7 to 43 +/- 8 ml/beat per m2 (P < 0.05). The heart size remained stable over the long term. Following cardiomyoplasty, the number of hospitalizations due to congestive heart failure was reduced to 0.4 hospitalizations/patient per year (preoperative: 2.5, P < 0.05). Computed tomography scans showed at long term a preserved latissimus dorsi muscle structure in 84% of patients. Survival probability at 7 years is 54%. Six patients underwent heart transplant after cardiomyoplasty (mean delay: 25 months), due to the natural evolution of their underlying heart disease. There were no specific technical difficulties. CONCLUSIONS: Clinically, this procedure reverses heart failure, improves functional class and ameliorates quality of life. The latissimus dorsi muscle histological structure is maintained at long-term, when postoperative electrostimulation is performed, avoiding excessive stimulation. Cardiomyoplasty may delay or prevent the progression of heart failure and the indication of cardiac transplantation.

Study of muscular and ventricular function in dynamic cardiomyoplasty: a ten-year follow-up.

BACKGROUND: The basic physiologic principle underlying cardiomyoplasty is long-term electrostimulation of a latissimus dorsi muscle (LDM) wrapped around the heart to obtain a phasic activity that could be integrated with ventricular kinetics. The aim of cardiomyoplasty is to prolong survival and to improve the quality of life of patients with severe chronic and irreversible myocardial failure by improving systolic contraction and correcting diastolic dysfunction. METHODS: To evaluate the long-term outcome of cardiomyoplasty, we investigated 82 patients electively undergoing this procedure in-our hospital. All patients had severe chronic heart failure that did not respond to optimal medical treatment. Patients had a mean age of 50 +/- 12 years (84% males). The cause of heart failure was ischemic (55%), idiopathic cardiomyopathy (34%), ventricular tumor (6%), and other (5%). The mean follow-up was 4.3 years. RESULTS: The mean New York Heart Association functional class improved after operation from 3.2 to 1.8. Average radioisotopic left ventricular ejection fraction increased from 17% +/- 6% to 28% +/- 3% (p < 0.05). Stroke volume index increased from 35 +/- 9 to 46 +/- 8 ml/beat/m2 (p < 0.05). The heart size remained stable at long term (evaluated by echo and computed tomography scanning). After cardiomyoplasty the number of successive hospitalizations resulting from congestive heart failure was reduced to 0.4 hospitalizations/patient/year (before operation 2.5, p < 0.05). Computed tomography scans showed at long-term a preserved LDM structure in 82% of patients who underwent operation. Survival probability at 7 years was 54% for the totality of patients, and 66% for patients who underwent operation in New York Heart Association functional class 3. Five patients underwent heart transplantation after cardiomyoplasty (mean delay 29 months), principally as a result of the natural evolution of their underlying heart disease, without major technical difficulties. CONCLUSIONS: Our 10-year clinical experience demonstrates that cardiomyoplasty increases ejection fraction, improves functional class, and ameliorates quality of life. Ventricular volumes and diameters remain stable long term. LDM structure is maintained long term if electrostimulation is performed avoiding excessive myostimulation. Patient selection is the most important determinant for early and late outcome. Late death in patients undergoing cardiomyoplasty is principally due to sudden death. Our future aim is to incorporate a cardioverter-defibrillator in the cardiomyostimulator, thus improving long-term results. Cardiomyoplasty may delay or prevent end-stage heart failure and the need for heart transplantation.

Directed TOCSY, a method for selection of directed correlations by optimal combinations of isotropic and longitudinal mixing.

The directed TOCSY pulse sequence element transfers coherence predominantly into "forward-directed" antiphase coherences while simultaneously suppressing in-phase and "backward-directed" antiphase coherences. This novel selection principle, based on the "direction" of the target coherences, provides a new approach for the simplification of crowded spectra. In this article, the theory of directed TOCSY is presented for linear spin systems that are frequently found in carbon-labeled biomolecules.

Bent helix formation between RNA hairpins with complementary loops.

The initial interaction between the ColE1 plasmid specific transcripts RNA I and RNA II, which function as antisense regulators of plasmid replication, comprises a transient complex between complementary loops found within the RNA secondary structures. Multidimensional heteronuclear magnetic resonance spectroscopy was used to characterize complexes formed between model RNA hairpins having seven nucleotide complementary loops. Seven base pairs are formed in the loop-loop helix, with continuous helical stacking of the loop residues on the 3' side of their helical stems. A sharp bend in the loop-loop helix, documented by gel electrophoresis, narrows the major groove and allows bridging of the phosphodiester backbones across the major groove in order to close the hairpin loops at their 5'-ends. The bend is further enhanced by the binding of Rom, a ColE1 encoded protein that regulates replication.

Sequential correlation of anomeric ribose protons and intervening phosphorus in RNA oligonucleotides by a 1H, 13C, 31P triple resonance experiment: HCP-CCH-TOCSY.

A three-dimensional 1H, 13C, 31P triple resonance experiment, HCP-CCH-TOCSY, is presented which provides unambiguous through-bond correlation of all 1H ribose protons on the 5' and 3' sides of the intervening phosphorus along the backbone bonding network in 13C-labeled RNA oligonucleotides. The correlation of the complete ribose spin system to the intervening phosphorus is obtained by adding a C,C-TOCSY coherence transfer step to the triple resonance HCP experiment. The C,C-TOCSY transfer step, which utilizes the large and relatively uniform 1J(C,C) coupling constant (approximately 40 Hz for ribose carbons), efficiently correlates the phosphorus-coupled carbons observed in the HCP correlation experiment (i.e., C4' and C5' in the 5' direction and C4' and C3' in the 3' direction) to all other carbons in the ribose spin system. Of the additional correlations observed in the HCP-CCH-TOCSY, that to the relatively well-resolved anomeric H1',C1' resonance pairs provides the greatest gain in terms of facilitating assignment. The gain in spectral resolution afforded by chemical shift labeling with the anomeric resonances should provide a more robust pathway for sequential assignment over the intervening phosphorus in larger RNA oligonucleotides. The HCP-CCH-TOCSY experiment is demonstrated on a uniformly 13C, 15N-labeled 19-nucleotide RNA stem-loop, derived from the antisense RNA I molecule found in the ColE1 plasmid replication control system.

Determination of a complete set of coupling constants in 13C-labeled oligonucleotides.

Three experiments are introduced to determine a complete set of coupling constants in RNA oligomers. In the HCCH-E.COSY experiment, the vicinal proton-proton coupling constants can be measured with high accuracy. In the P-FIDS-CT-HSQC experiment, vicinal proton-phosphorus and carbon-phosphorus couplings are measured that depend on the phosphodiester backbone torsion angles beta and epsilon. In the refocussed HMBC experiment, vicinal carbon-proton couplings are measured that depend on the glycosidic torsion angle chi.