Measuring “Unmeasurable” Folding Kinetics of Proteins by Single-Molecule Force Spectroscopy

Folding and unfolding are fundamental biological processes in cell and are important for the biological functions of proteins. Characterizing the folding and unfolding kinetics of proteins is important for understanding the energetic landscape leading to the active native conformations of these molecules. However, the thermal or chemical-induced unfolding of many proteins is irreversible in vitro, precluding characterization of the folding kinetics of such proteins, just as it is impossible to “un-boil” an egg. Irreversible unfolding often manifests as irreversible aggregation of unfolded polypeptide chains, which typically occurs between denatured protein molecules in response to the exposure of hydrophobic residues to solvent. An example of such a protein where thermal denaturation results in irreversible aggregation is the β-1,4 endoxylanase from Bacillus circulans (BCX). Here, we report the use of single-molecule atomic force microscopy to directly measure the folding kinetics of BCX in vitro. By mechanically unfolding BCX, we essentially allowed only one unfolded molecule to exist in solution at a given time, effectively eliminating the possibility for aggregation. We found that BCX can readily refold back to the native state, allowing us to measure its folding kinetics for the first time. Our results demonstrate that single-molecule force-spectroscopy-based methods can adequately tackle the challenge of “un-boiling eggs”, providing a general methodology to characterize the folding kinetics of many proteins that suffer from irreversible denaturation and thus cannot be characterized using traditional equilibrium methodologies.     

Nanomechanical Properties of Tenascin-X Revealed by Single-Molecule Force Spectroscopy

Tenascin-X is an extracellular matrix protein and binds a variety of molecules in extracellular matrix and on cell membrane. Tenascin-X plays important roles in regulating the structure and mechanical properties of connective tissues. Using single-molecule atomic force microscopy, we have investigated the mechanical properties of bovine tenascin-X in detail. Our results indicated that tenascin-X is an elastic protein and the fibronectin type III (FnIII) domains can unfold under a stretching force and refold to regain their mechanical stability upon the removal of the stretching force. All the 30 FnIII domains of tenascin-X show similar mechanical stability, mechanical unfolding kinetics, and contour length increment upon domain unfolding, despite their large sequence diversity. In contrast to the homogeneity in their mechanical unfolding behaviors, FnIII domains fold at different rates. Using the 10th FnIII domain of tenascin-X (TNXfn10) as a model system, we constructed a polyprotein chimera composed of alternating TNXfn10 and GB1 domains and used atomic force microscopy to confirm that the mechanical properties of TNXfn10 are consistent with those of the FnIII domains of tenascin-X. These results lay the foundation to further study the mechanical properties of individual FnIII domains and establish the relationship between point mutations and mechanical phenotypic effect on tenascin-X. Moreover, our results provided the opportunity to compare the mechanical properties and design of different forms of tenascins. The comparison between tenascin-X and tenascin-C revealed interesting common as well as distinguishing features for mechanical unfolding and folding of tenascin-C and tenascin-X and will open up new avenues to investigate the mechanical functions and architectural design of different forms of tenascins.     

Single-Molecule and FRET Fluorescence Correlation Spectroscopy Analyses of Phage DNA Packaging: Colocalization of Packaged Phage T4 DNA Ends within the Capsid

Linear DNAs of any sequence can be packaged into empty viral procapsids by the phage T4 terminase with high efficiency in vitro. Packaging substrates of 5 kbp and 50 kbp, terminated by energy transfer dye pairs, were constructed from plasmid and λ phage DNAs. Nuclease and fluorescence correlation spectroscopy (FCS) assays showed that ∼ 20% of the substrate DNA was packaged and that the DNA dye ends of the packaged DNA were protected from nuclease digestion. Upon packaging, both 5-kbp and  50-kbp DNAs produced comparable fluorescence resonance energy transfer (FRET) between Cy5 and Cy5.5 double-dye terminated DNAs. Single-molecule FRET (sm-FRET) and photobleaching analysis shows that FRET is intramolecular rather than intermolecular upon packaging of most procapsids and demonstrates that single-molecule detection allows mechanistic analysis of packaging in vitro. FRET-FCS and sm-FRET measurements are comparable and show that both the 5-kbp and the  50-kbp packaged DNA ends are held within 8-9 nm of each other, within the dimensions of the long axis of the procapsid portal. The calculated distribution of FRET distances is relatively narrow for both FRET-FCS and sm-FRET, suggesting that the two packaged DNA ends are held at the same fixed distance relative to each other in most capsids. Because one DNA end is known to be positioned for ejection through the portal, it can be inferred that both DNAs ends are held in proximity to the portal entrance and ejection channel. The analysis suggests that a DNA loop, rather than a DNA end, is translocated by the packaging motor to fill the procapsid.     

Sensory rhodopsin II/transducer complex formation in detergent and in lipid bilayers studied with FRET

The photophobic receptor from Natronomonas pharaonis (NpSRII) forms a photo-signalling complex with its cognate transducer (NpHtrII). In order to elucidate the complex formation in more detail, we have studied the intermolecular binding of both constituents (NpSRII and NpHtrII₁₅₇; truncated at residue 157) in detergent buffers, and in lipid bilayers using FRET. The data for hetero-dimer formation of NpSRII/NpHtrII in detergent agrees well with K_D values (∼ 200 nM) described in the literature. In lipid bilayers, the binding affinity between proteins in the NpSRII/NpHtrII complex is at least one order of magnitude stronger. In detergent the strength of binding is similar for both homo-dimers (NpSRII/NpSRII and NpHtrII/NpHtrII) but significantly weaker (K_D  ∼ 16 μM) when compared to the hetero-dimer. The intermolecular binding is again considerably stronger in lipid bilayers; however, it is not as strong as that observed for the hetero-dimer. At a molar transducer/lipid ratio of 1:2000, which is still well above physiological concentrations, only 40% homo-dimers are formed. Apparently, in cell membranes the formation of the assumed functionally active oligomeric 2:2 complex depends on the full-length transducer including the helical cytoplasmic part, which is thought to tighten the transducer-dimer association.     

Integrin αIIbβ3 in a Membrane Environment Remains the Same Height after Mn Activation when Observed by Cryoelectron Tomography

Integrins perform the critical function of signalling cell attachment to the extracellular matrix or to other cells. This signalling is done through a structural change propagated bidirectionally across the plasma membrane. Integrin activation has been extensively studied with ectodomain constructs, but the structural change within intact, membrane-bound molecules remains a subject of live debate. Using cryoelectron tomography, we examined the simplest predication of the different integrin activation models, i.e., the change in height of the molecules. Analysis using techniques that compensate for the missing wedge during alignment and averaging and that search for patterns in the structure of the aligned molecular subvolumes extracted from the tomogram reveals that the vast majority of molecules show no dramatic height change upon Mn²⁺-induced activation of membrane-bound integrins when compared with an inactive integrin control group. Thus, the result is inconsistent with the switchblade activation model.     

Structural Evidence for a Sequential Release Mechanism for Activation of Heterotrimeric G Proteins

Heptahelical G-protein (heterotrimeric guanine nucleotide-binding protein)-coupled receptors couple to heterotrimeric G proteins to relay extracellular signals to intracellular signaling networks, but the molecular mechanism underlying guanosine 5′-diphosphate (GDP) release by the G protein α-subunit is not well understood. Amino acid substitutions in the conserved α5 helix of G_i, which extends from the C-terminal region to the nucleotide-binding pocket, cause dramatic increases in basal (receptor-independent) GDP release rates. For example, mutant Gα_i1-T329A shows an 18-fold increase in basal GDP release rate and, when expressed in culture, it causes a significant decrease in forskolin-stimulated cAMP accumulation. The crystal structure of Gα_i1-T329A·GDP shows substantial conformational rearrangement of the switch I region and additional striking alterations of side chains lining the catalytic pocket that disrupt the Mg⁺² coordination sphere and dislodge bound Mg⁺². We propose a “sequential release” mechanism whereby a transient conformational change in the α5 helix alters switch I to induce GDP release. Interestingly, this mechanistic model for heterotrimeric G protein activation is similar to that suggested for the activation of the plant small G protein Rop4 by RopGEF8.     

Elongation of amyloid fibrils through lateral binding of monomers revealed by total internal reflection fluorescence microscopy

Amyloid fibrils are fibrillar aggregates of denatured proteins associated with a large number of amyloidoses. The formation of amyloid fibrils has been considered to occur by nucleation and elongation. Real-time imaging of the elongation as well as linear morphology of amyloid fibrils suggests that all elongation events occur at the growing ends of fibrils. On the other hand, we suggested that monomers also bind to the lateral sides of preformed fibrils during the seed-dependent elongation, diffuse to the growing ends, and finally make further conformation changes to the mature amyloid fibrils. To examine lateral binding during the elongation of fibrils, we used islet amyloid polypeptide (IAPP), which has been associated with type II diabetes, and prepared IAPP modified with the fluorescence dye, Alexa532. By monitoring the elongation process with amyloid specific thioflavin T and Alexa532 fluorescence, we obtained overlapping images of the two fluorescence probes, which indicated lateral binding. These results are similar to the surface diffusion-dependent growth of crystals, further supporting the similarities between amyloid fibrillation and the crystallization of substances.     

Structural Information, Resolution, and Noise in High-Resolution Atomic Force Microscopy Topographs

AFM has developed into a powerful tool in structural biology, providing topographs of proteins under close-to-native conditions and featuring an outstanding signal/noise ratio. However, the imaging mechanism exhibits particularities: fast and slow scan axis represent two independent image acquisition axes. Additionally, unknown tip geometry and tip-sample interaction render the contrast transfer function nondefinable. Hence, the interpretation of AFM topographs remained difficult. How can noise and distortions present in AFM images be quantified? How does the number of molecule topographs merged influence the structural information provided by averages? What is the resolution of topographs? Here, we find that in high-resolution AFM topographs, many molecule images are only slightly disturbed by noise, distortions, and tip-sample interactions. To identify these high-quality particles, we propose a selection criterion based on the internal symmetry of the imaged protein. We introduce a novel feature-based resolution analysis and show that AFM topographs of different proteins contain structural information beginning at different resolution thresholds: 10 Å (AqpZ), 12 Å (AQP0), 13 Å (AQP2), and 20 Å (light-harvesting-complex-2). Importantly, we highlight that the best single-molecule images are more accurate molecular representations than ensemble averages, because averaging downsizes the z-dimension and “blurs” structural details.Abbreviations: 2D, two-dimensional; 3D, three-dimensional; ACV, auto-correlation value; AFM, atomic force microscopy; AQP0, aquaporin-0; AQP2, aquaporin-2; AqpZ, aquaporin-Z; bR, bacteriorhodopsin; CCV, cross-correlation value; CTF, contrast transfer function; DPR, differential phase residual; EM, electron microscopy; FRC, Fourier ring correlation; FSC, Fourier shell correlation; IS, internal symmetry; LH2, light-harvesting-complex 2; RMSD, root mean-square deviation; SD, standard deviation; SNR, signal/noise ratio; SSNR, spectral signal/noise ratio     

At Low Concentrations, 3,4-Dihydroxyphenylacetic Acid (DOPAC) Binds Non-Covalently to α-Synuclein and Prevents Its Fibrillation

Several studies have shown that catecholamines can inhibit the fibrillation of α-synuclein (α-Syn), a small presynaptic protein whose aggregation is believed to be a critical step in the etiology of Parkinson's disease and several other neurodegenerative disorders. However, the mechanism of this inhibition is uncertain. We show here that substoichiometric concentrations of 3,4-dihydroxyphenylacetic acid (DOPAC), a normal product of the metabolism of dopamine, can inhibit the fibrillation of α-Syn, due to non-covalent binding of DOPAC to α-Syn monomer. Intriguingly, the presence of α-Syn accelerates the spontaneous oxidation of DOPAC, and the oxidized form of DOPAC (the quinone) is responsible for the fibrillation inhibition. In addition, the presence of DOPAC leads to the oxidation of the methionine residues of α-Syn, probably due to the H₂O₂ production as a by-product of DOPAC oxidation. The lack of fibrillation results from the formation of stable oligomers, which are very similar to those observed transiently at early stages of the α-Syn fibrillation. A possible explanation for this phenomenon is that DOPAC stabilizes the normally transient oligomers and prevents them from subsequent fibril formation. The analysis of the α-Syn Y39W variant suggests that DOPAC binds non-covalently to the same N-terminal region of α-Syn as lipid vesicles, probably in the vicinity of residue 39. In contrast to the compounds with 1,2-dihydroxyphenyl groups (DOPAC and catechol), their 1,4-dihydroxyphenyl isomers (hydroquinone and homogentisic acid) are able to modify α-Syn covalently, probably due to the less steric hindrance in the Michael addition.     

Nucleotide-Induced Flexibility Change in Neck Linkers of Dimeric Kinesin as Detected by Distance Measurements Using Spin-Labeling EPR

Using dipolar continuous-wave and pulsed electron paramagnetic resonance methods, we have determined the distribution of the distances between two spin labels placed on the middle of each of the neck linkers of dimeric kinesin. In the absence of microtubules, the distance was centered at 3.3 nm, but displayed a broad distribution with a width of 2.7 nm. This broad distribution implies that the linkers are random coils and extend well beyond the 2.5-nm distance expected of crystal structures. In the presence of microtubules, two linker populations were found: one similar to that observed in the absence of microtubules (a broad distribution centered at 3.3 nm), and the second population with a narrower distribution centered at 1.3-2.5 nm. In the absence of nucleotide but in the presence of microtubules, ∼ 40% of the linkers were at a distance centered at 1.9 nm with a 1.2-nm width; the remaining fraction was at 3.3 nm, as before. This suggests that neck linkers exhibit dynamics covering a wide distance range between 1.0 and 5.0 nm. In the presence of ATP analogs adenosine 5′-(β,γ-imido)triphosphate and adenosine 5′-(γ-thio)triphosphate, 40-50% of the spins showed a very narrow distribution centered at 1.6 nm, with a width of 0.4-0.5 nm. The remaining population displayed the broad 3.3-nm distribution. Under these conditions, a large fraction of linkers are docked firmly onto a motor core or microtubule, while the remainder is disordered.We propose that large nucleotide-dependent flexibility changes in the linkers contribute to the directional bias of the kinesin molecule stepping 8 nm along the microtubule.     

Stepwise Transition of the Tetra-Manganese Complex of Photosystem II to a Binuclear Mn2(μ-O)2 Complex in Response to a Temperature Jump: A Time-Resolved Structural Investigation Employing X-Ray Absorption Spectroscopy

In oxygenic photosynthesis, water is oxidized at a protein-cofactor complex comprising four Mn atoms and, presumably, one calcium. Using multilayers of Photosystem II membrane particles, we investigated the time course of the disassembly of the Mn complex initiated by a temperature jump from 25°C to 47°C and terminated by rapid cooling after distinct heating periods. We monitored polarographically the oxygen-evolution activity, the amount of the Y_D^ox radical and of released Mn²⁺ by EPR spectroscopy, and the structure of the Mn complex by x-ray absorption spectroscopy (XAS, EXAFS). Using a novel approach to analyze time-resolved EXAFS data, we identify three distinct phases of the disassembly process: (1) Loss of the oxygen-evolution activity and reduction of Y_D^ox occur simultaneously (k₁ = 1.0 min⁻¹). EXAFS spectra reveal the concomitant loss of an absorber-backscatterer interaction between heavy atoms separated by ~3.3 Å, possibly related to Ca release. (2) Subsequently, two Mn(III) or Mn(IV) ions seemingly separated by ~2.7 Å in the native complex are reduced to Mn(II) and released (k₂ = 0.18 min⁻¹). The x-ray absorption spectroscopy data is highly suggestive that the two unreleased Mn ions form a di-μ-oxo bridged Mn(III)₂ complex. (3) Finally, the tightly-bound Mn₂(μ-O)₂ unit is slowly reduced and released (k₃ = 0.014 min⁻¹).     

Mechanism divergence in microbial rhodopsins

A fundamental design principle of microbial rhodopsins is that they share the same basic light-induced conversion between two conformers. Alternate access of the Schiff base to the outside and to the cytoplasm in the outwardly open “E” conformer and cytoplasmically open “C” conformer, respectively, combined with appropriate timing of pKa changes controlling Schiff base proton release and uptake make the proton path through the pumps vectorial. Phototaxis receptors in prokaryotes, sensory rhodopsins I and II, have evolved new chemical processes not found in their proton pump ancestors, to alter the consequences of the conformational change or modify the change itself. Like proton pumps, sensory rhodopsin II undergoes a photoinduced E → C transition, with the C conformer a transient intermediate in the photocycle. In contrast, one light-sensor (sensory rhodopsin I bound to its transducer HtrI) exists in the dark as the C conformer and undergoes a light-induced C → E transition, with the E conformer a transient photocycle intermediate. Current results indicate that algal phototaxis receptors channelrhodopsins undergo redirected Schiff base proton transfers and a modified E → C transition which, contrary to the proton pumps and other sensory rhodopsins, is not accompanied by the closure of the external half-channel. The article will review our current understanding of how the shared basic structure and chemistry of microbial rhodopsins have been modified during evolution to create diverse molecular functions: light-driven ion transport and photosensory signaling by protein–protein interaction and light-gated ion channel activity.     

Single-molecule imaging and manipulation of biomolecular machines and systems

Background

Biological molecular machines support various activities and behaviors of cells, such as energy production, signal transduction, growth, differentiation, and migration.

Activation of phospholipase A2 by 1-palmitoyl-2-(9'-oxo-nonanoyl)-sn-glycero-3-phosphocholine in vitro

Oxidative stress leads to drastic modifications of both the biophysical properties of biomembranes and their associated chemistry imparted upon the formation of oxidatively modified lipids. To this end, oxidized phospholipid derivatives bearing an aldehyde function, such as 1-palmitoyl-2-(9'-oxo-nonanoyl)-sn-glycero-3-phosphocholine (PoxnoPC) can covalently react with proteins that come into direct contact. Intriguingly, we observed PoxnoPC in a 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) matrix to shorten and abolish the lag time in the action of phospholipase A2 (PLA2) on this composite substrate, with concomitant augmented decrement in pH, indicating more extensive hydrolysis, which was in keeping with enhanced 90° light scattering. The latter was abolished by the aldehyde scavenger methoxyamine, thus suggesting the involvement of Schiff base. Enhanced hydrolysis of a fluorescent phospholipid analogue was seen for PLA2 preincubated with PoxnoPC. Mixing PLA2 with submicellar (22 µM) PoxnoPC caused a pronounced increase in Thioflavin T fluorescence, in keeping with the formation of amyloid-type fibers, which were seen also by electron microscopy.     

Site-directed mutagenesis on the heme axial-ligands of cytochrome b559 in photosystem II by using cyanobacteria Synechocystis PCC 6803

Cytochrome (cyt) b559 has been proposed to play an important role in the cyclic electron flow processes that protect photosystem II (PSII) from light-induced damage during photoinhibitory conditions. However, the exact role(s) of cyt b559 in the cyclic electron transfer pathway(s) in PSII remains unclear. To study the exact role(s) of cyt b559, we have constructed a series of site-directed mutants, each carrying a single amino acid substitution of one of the heme axial-ligands, in the cyanobacterium Synechocystis sp. PCC6803. In these mutants, His-22 of the α or the β subunit of cyt b559 was replaced with either Met, Glu, Tyr, Lys, Arg, Cys or Gln. On the basis of oxygen-evolution and chlorophyll a fluorescence measurements, we found that, among all mutants that were constructed, only the H22Kα mutant grew photoautotrophically, and accumulated stable PSII reaction centers (∼ 81% compared to wild-type cells). In addition, we isolated one pseudorevertant of the H22Yβ mutant that regained the ability to grow photoautotrophically and to assemble stable PSII reaction centers (∼ 79% compared to wild-type cells). On the basis of 77 K fluorescence emission measurements, we found that energy transfer from the phycobilisomes to PSII reaction centers was uncoupled in those cyt b559 mutants that assembled little or no stable PSII. Furthermore, on the basis of immunoblot analyses, we found that in thylakoid membranes of cyt b559 mutants that assembled little or no PSII, the amounts of the D1, D2, cyt b559α and β polypeptides were very low or undetectable but their CP47 and PsaC polypeptides were accumulated to the wild-type level. We also found that the amounts of cyt b559β polypeptide were significantly increased (larger than two folds) in thylakoid membranes of cyt b559 H22YβPS+ mutant cells. We suspected that the increase in the amounts of cyt b559 H22YβPS+ mutant polypeptides in thylakoid membranes might facilitate the assembly of functional PSII in cyt b559 H22YβPS+ mutant cells. Moreover, we found that isolated His-tagged PSII particles from H22Kα mutant cells gave rise to redox-induced optical absorption difference spectra of cyt b559. Therefore, our results concluded that significant fractions of H22Kα mutant PSII particles retained the heme of cyt b559. Finally, this work is the first report of cyt b559 mutants having substitutions of an axial heme-ligands that retain the ability to grow photoautotrophically and to assemble stable PSII reaction centers. These two cyt b559 mutants (H22Kα and H22YβPS+) and their PSII reaction centers will be very suitable for further biophysical and biochemical studies of the functional role(s) of cyt b559 in PSII.     

Structural elucidation of transmembrane transporter protein bilitranslocase: Conformational analysis of the second transmembrane region TM2 by molecular dynamics and NMR spectroscopy

Membrane proteins represent about a third of the gene products in most organisms, as revealed by the genome sequencing projects. They account for up to two thirds of known drugable targets, which emphasizes their critical pharmaceutical importance. Here we present a study on bilitranslocase (BTL) (TCDB 2.A.65), a membrane protein primarily involved in the transport of bilirubin from blood to liver cells. Bilitranslocase has also been identified as a potential membrane transporter for cellular uptake of several drugs and due to its implication in drug uptake, it is extremely important to advance the knowledge about its 3D structure. However, at present, only a limited knowledge is available beyond the primary structure of BTL. It has been recently confirmed experimentally that one of the four computationally predicted transmembrane segments of bilitranslocase, TM3, has a helical structure with hydrophilic amino acid residues oriented towards one side, which is typical for transmembrane domains of membrane proteins. In this study we confirmed by the use of multidimensional NMR spectroscopy that the second transmembrane segment, TM2, also appears in a form of α-helix. The stability of this polypeptide chain was verified by molecular dynamics (MD) simulation in dipalmitoyl phosphatidyl choline (DPPC) and in sodium dodecyl sulfate (SDS) micelles. The two α-helices, TM2 corroborated in this study, and TM3 confirmed in our previous investigation, provide reasonable building blocks of a potential transmembrane channel for transport of bilirubin and small hydrophilic molecules, including pharmaceutically active compounds.     

Evidence for an interaction between the SH3 domain and the N-terminal extension of the essential light chain in class II myosins

The function of the src-homology 3 (SH3) domain in class II myosins, a distinct beta-barrel structure, remains unknown. Here, we provide evidence, using electron cryomicroscopy, in conjunction with light-scattering, fluorescence and kinetic analyses, that the SH3 domain facilitates the binding of the N-terminal extension of the essential light chain isoform (ELC-1) to actin. The 41 residue extension contains four conserved lysine residues followed by a repeating sequence of seven Pro/Ala residues. It is widely believed that the highly charged region interacts with actin, while the Pro/Ala-rich sequence forms a rigid tether that bridges the approximately 9 nm distance between the myosin lever arm and the thin filament. In order to localize the N terminus of ELC in the actomyosin complex, an engineered Cys was reacted with undecagold-maleimide, and the labeled ELC was exchanged into myosin subfragment-1 (S1). Electron cryomicroscopy of S1-bound actin filaments, together with computer-based docking of the skeletal S1 crystal structure into 3D reconstructions, showed a well-defined peak for the gold cluster near the SH3 domain. Given that SH3 domains are known to bind proline-rich ligands, we suggest that the N-terminal extension of ELC interacts with actin and modulates myosin kinetics by binding to the SH3 domain during the ATPase cycle.     

Multifaceted applications of nanomaterials in cell engineering and therapy

Nanomaterials with superior physiochemical properties have been rapidly developed and integrated in every aspect of cell engineering and therapy for translating their great promise to clinical success. Here we demonstrate the multifaceted roles played by innovatively-designed nanomaterials in addressing key challenges in cell engineering and therapy such as cell isolation from heterogeneous cell population, cell instruction in vitro to enable desired functionalities, and targeted cell delivery to therapeutic sites for prompting tissue repair. The emerging trends in this interdisciplinary and dynamic field are also highlighted, where the nanomaterial-engineered cells constitute the basis for establishing in vitro disease model; and nanomaterial-based in situ cell engineering are accomplished directly within the native tissue in vivo. We will witness the increasing importance of nanomaterials in revolutionizing the concept and toolset of cell engineering and therapy which will enrich our scientific understanding of diseases and ultimately fulfill the therapeutic demand in clinical medicine.