Alkaline pH-dependent differential unfolding characteristics of mesophilic and thermophilic homologs of dimeric serine hydroxymethyltransferase

Environmental variables such as pH can significantly influence the folding and stability of a protein molecule. In the present investigation, we compared the alkaline pH-induced unfolding of two homologous serine hydroxymethyltransferase from mesophilic Bacillus subtilis (bsSHMT) and thermophilic Bacillus stearothermophilus (bstSHMT) using various biophysical techniques. The thermophilic enzyme bstSHMT was found to be more resistant to alkaline denaturation compared to its mesophilic counterpart, bsSHMT. Unfolding studies using domain-swapped chimera, constructed by swapping the C-terminal domain of these two wild-type proteins, revealed that C-terminal domain plays a pivotal role in the folding, stability and subunit interaction of these proteins. Primary amino acid sequence analysis of the proteins showed that bsSHMT has six unconserved lysine residues in C-terminal domain, which are absent in bstSHMT. Chemical modification of lysine side chains resulted in stabilization of monomers, only in case of bsSHMT. Moreover, comparison between homology model of bsSHMT with the crystal structure of bstSHMT revealed that a small stretch of 11 amino acids at the end of C-terminal domain was found protruding outside the molecule as a flexible coiled structure in bsSHMT. Taken together these findings suggest that possibly the presence of these non-identical lysine moieties and a small extension of C-terminal domain may be responsible for low stability of bsSHMT under alkaline pH condition.     

Structural and Functional Studies of gpX of Escherichia coli Phage P2 Reveal a Widespread Role for LysM Domains in the Baseplates of Contractile-Tailed Phages

A variety of bacterial pathogenicity determinants, including the type VI secretion system and the virulence cassettes from Photorhabdus and Serratia, share an evolutionary origin with contractile-tailed myophages. The well-characterized Escherichia coli phage P2 provides an excellent system for studies related to these systems, as its protein composition appears to represent the “minimal” myophage tail. In this study, we used nuclear magnetic resonance (NMR) spectroscopy to determine the solution structure of gpX, a 68-residue tail baseplate protein. Although the sequence and structure of gpX are similar to those of LysM domains, which are a large family associated with peptidoglycan binding, we did not detect a peptidoglycan-binding activity for gpX. However, bioinformatic analysis revealed that half of all myophages, including all that possess phage T4-like baseplates, encode a tail protein with a LysM-like domain, emphasizing a widespread role for this domain in baseplate function. While phage P2 gpX comprises only a single LysM domain, many myophages display LysM domain fusions with other tail proteins, such as the DNA circulation protein found in Mu-like phages and gp53 of T4-like phages. Electron microscopy of P2 phage particles with an incorporated gpX-maltose binding protein fusion revealed that gpX is located at the top of the baseplate, near the junction of the baseplate and tail tube. gpW, the orthologue of phage T4 gp25, was also found to localize to this region. A general colocalization of LysM-like domains and gpW homologues in diverse phages is supported by our bioinformatic analysis.     

HU histone-like DNA-binding protein from <Emphasis Type="Italic">Thermus thermophilus</Emphasis>: structural and evolutionary analyses

The histone-like DNA-binding proteins (HU) serve as model molecules for protein thermostability studies, as they function in different bacteria that grow in a wide range of temperatures and show sequence diversity under a common fold. In this work, we report the cloning of the hutth gene from Thermus thermophilus, the purification and crystallization of the recombinant HUTth protein, as well as its X-ray structure determination at 1.7 Å. Detailed structural and thermodynamic analyses were performed towards the understanding of the thermostability mechanism. The interaction of HUTth protein with plasmid DNA in solution has been determined for the first time with MST. Sequence conservation of an exclusively thermophilic order like Thermales, when compared to a predominantly mesophilic order (Deinococcales), should be subject, to some extent, to thermostability-related evolutionary pressure. This hypothesis was used to guide our bioinformatics and evolutionary studies. We discuss the impact of thermostability adaptation on the structure of HU proteins, based on the detailed evolutionary analysis of the Deinococcus–Thermus phylum, where HUTth belongs. Furthermore, we propose a novel method of engineering thermostable proteins, by combining consensus-based design with ancestral sequence reconstruction. Finally, through the structure of HUTth, we are able to examine the validity of these predictions. Our approach represents a significant advancement, as it explores for the first time the potential of ancestral sequence reconstruction in the divergence between a thermophilic and a mainly mesophilic taxon, combined with consensus-based engineering.     

Amino Acid Substitutions in Cold-Adapted Proteins from Halorubrum lacusprofundi,an Extremely Halophilic Microbe from Antarctica

The halophilic Archaeon Halorubrum lacusprofundi, isolated from the perennially cold and hypersaline Deep Lake in Antarctica, was recently sequenced and compared to 12 Haloarchaea from temperate climates by comparative genomics. Amino acid substitutions for 604 H. lacusprofundi proteins belonging to conserved haloarchaeal orthologous groups (cHOGs) were determined and found to occur at 7.85% of positions invariant in proteins from mesophilic Haloarchaea. The following substitutions were observed most frequently: (a) glutamic acid with aspartic acid or alanine; (b) small polar residues with other small polar or non-polar amino acids; (c) small non-polar residues with other small non-polar residues; (d) aromatic residues, especially tryptophan, with other aromatic residues; and (e) some larger polar residues with other similar residues. Amino acid substitutions for a cold-active H. lacusprofundi β-galactosidase were then examined in the context of a homology modeled structure at residues invariant in homologous enzymes from mesophilic Haloarchaea. Similar substitutions were observed as in the genome-wide approach, with the surface accessible regions of β-galactosidase displaying reduced acidity and increased hydrophobicity, and internal regions displaying mainly subtle changes among smaller non-polar and polar residues. These findings are consistent with H. lacusprofundi proteins displaying amino acid substitutions that increase structural flexibility and protein function at low temperature. We discuss the likely mechanisms of protein adaptation to a cold, hypersaline environment on Earth, with possible relevance to life elsewhere.     

"Cold spots" in protein cold adaptation: Insights from normalized atomic displacement parameters (B'-factors)

Many analyses published in the last decade suggest that enzymes isolated from cold-adapted organisms are characterized by a higher flexibility of their molecular structure. Recently, it has been argued that all cold-adapted enzymes with catalytic efficiency greater than that of their mesophilic counterparts display local flexibility or rigidity that are likely to cooperate, each acting on specific areas of the enzyme structure. Here we report an analysis of the normalized thermal B-factor distributions in psychrophilic proteins compared with those of their mesophilic and thermophilic counterparts with the aim to detect statistically significant local variations of relative backbone flexibility possibly linked to cold adaptation. We utilized a strategy based mainly on intra-family comparison of local distribution of normalized B-factors. After careful statistical treatment of data, the picture emerging from our results suggests that the distribution of the flexibility in psychrophilic enzymes is locally more heterogeneous than in their respective mesophilic homologues.     

A Comparative Molecular Dynamics Study of Thermophilic and Mesophilic Ribonuclease HI Enzymes

Abstract

We studied a pair of homologous thermophilic and mesophilic ribonuclease HI enzymes by molecular dynamics simulations. Each protein was subjected to three 5 ns simulations in explicit water at both 310 K and 340 K. The thermophilic enzyme showed larger overall positional fluctuations at both temperatures, while only the mesophilic enzyme at the higher temperature showed significant instability. When the temperature is changed, the relative flexibility of different local segments on the two proteins changed differently. Principal component analysis showed that the simulations of the two proteins explored largely overlapping regions in the conformational space. However, at 340 K, the collective structure variations of the thermophilic protein are different from those of the mesophilic protein. Our results, although not in accordance with the view that hyperthermostability of proteins may originate from their conformational rigidity, are consistent with several recent experimental and simulation studies which showed that thermophilic proteins may be conformationally more flexible than their mesophilic counterparts. The decorrelation between conformational rigidity and hyperthermostability may be attributed to the temperature dependence and long range nature of electrostatic interactions that play more important roles in the structural stability of thermophilic proteins.     

ER stress-mediated apoptosis induced by celastrol in cancer cells and important role of glycogen synthase kinase-3β in the signal network

HeLa cells treated with celastrol, a natural compound with inhibitive effect on proteasome, exhibited increase in apoptotic rate and characteristics of apoptosis. To clarify the signal network activated by celastrol to induce apoptosis, both the direct target proteins and undirect target proteins of celastrol were searched in the present study. Proteasome catalytic subunit β1 was predicted by computational analysis to be a possible direct target of celastrol and confirmed by checking direct effect of celastrol on the activity of recombinant human proteasome subunit β1 in vitro. Undirect target-related proteins of celastrol were searched using proteomic studies including two-dimensional electrophoresis (2-DE) analysis and iTRAQ-based LC-MS analysis. Possible target-related proteins of celastrol such as endoplasmic reticulum protein 29 (ERP29) and mitochondrial import receptor Tom22 (TOM22) were found by 2-DE analysis of total cellular protein expression profiles. Further study showed that celastrol induced ER stress and ER stress inhibitor could ameliorate cell death induced by celastrol. Celastrol induced translocation of Bax into the mitochondria, which might be related to the upregulation of BH-3-only proteins such as BIM and the increase in the expression level of TOM22. To further search possible target-related proteins of celastrol in ER and ER-related fractions, iTRAQ-based LC-MS method was use to analyze protein expression profiles of ER/microsomal vesicles-riched fraction of cells with or without celastrol treatment. Based on possible target-related proteins found in both 2-DE analysis and iTRAQ-based LC-MS analysis, protein–protein interaction (PPI) network was established using bioinformatic analysis. The important role of glycogen synthase kinase-3β (GSK3β) in the signal cascades of celastrol was suggested. Pretreatment of LiCL, an inhibitor of GSK3β, could significantly ameliorate apoptosis induced by celastrol. On the basis of the results of the present study, possible signal network of celastrol activated by celastrol leading to apoptosis was predicted.     

Proteins maintain hydration at high [KCl] concentration regardless of content in acidic amino acids

Proteins of halophilic organisms, which accumulate molar concentrations of KCl in their cytoplasm, have a much higher content in acidic amino acids than proteins of mesophilic organisms. It has been proposed that this excess is necessary to maintain proteins hydrated in an environment with low water activity, either via direct interactions between water and the carboxylate groups of acidic amino acids or via cooperative interactions between acidic amino acids and hydrated cations. Our simulation study of five halophilic proteins and five mesophilic counterparts does not support either possibility. The simulations use the AMBER ff14SB force field with newly optimized Lennard-Jones parameters for the interactions between carboxylate groups and potassium ions. We find that proteins with a larger fraction of acidic amino acids indeed have higher hydration levels, as measured by the concentration of water in their hydration shell and the number of water/protein hydrogen bonds. However, the hydration level of each protein is identical at low (b_KCl = 0.15 mol/kg) and high (b_KCl = 2 mol/kg) KCl concentrations; excess acidic amino acids are clearly not necessary to maintain proteins hydrated at high salt concentration. It has also been proposed that cooperative interactions between acidic amino acids in halophilic proteins and hydrated cations stabilize the folded protein structure and would lead to slower dynamics of the solvation shell. We find that the translational dynamics of the solvation shell is barely distinguishable between halophilic and mesophilic proteins; if such a cooperative effect exists, it does not have that entropic signature.     

Thermostability of In Vitro Evolved Bacillus subtilis Lipase A: A Network and Dynamics Perspective

Proteins in thermophilic organisms remain stable and function optimally at high temperatures. Owing to their important applicability in many industrial processes, such thermostable proteins have been studied extensively, and several structural factors attributed to their enhanced stability. How these factors render the emergent property of thermostability to proteins, even in situations where no significant changes occur in their three-dimensional structures in comparison to their mesophilic counter-parts, has remained an intriguing question. In this study we treat Lipase A from Bacillus subtilis and its six thermostable mutants in a unified manner and address the problem with a combined complex network-based analysis and molecular dynamic studies to find commonality in their properties. The Protein Contact Networks (PCN) of the wild-type and six mutant Lipase A structures developed at a mesoscopic scale were analyzed at global network and local node (residue) level using network parameters and community structure analysis. The comparative PCN analysis of all proteins pointed towards important role of specific residues in the enhanced thermostability. Network analysis results were corroborated with finer-scale molecular dynamics simulations at both room and high temperatures. Our results show that this combined approach at two scales can uncover small but important changes in the local conformations that add up to stabilize the protein structure in thermostable mutants, even when overall conformation differences among them are negligible. Our analysis not only supports the experimentally determined stabilizing factors, but also unveils the important role of contacts, distributed throughout the protein, that lead to thermostability. We propose that this combined mesoscopic-network and fine-grained molecular dynamics approach is a convenient and useful scheme not only to study allosteric changes leading to protein stability in the face of negligible over-all conformational changes due to mutations, but also in other molecular networks where change in function does not accompany significant change in the network structure.     

Small multidrug resistance proteins: A multidrug transporter family that continues to grow

The small multidrug resistance (SMR) protein family is a bacterial multidrug transporter family. As suggested by their title, SMR proteins are composed of four transmembrane α-helices of approximately 100-140 amino acids in length. Since their designation as a family, many homologues have been identified and characterized both structurally and functionally. In this review the topology, structure, drug resistance, drug binding, and transport mechanisms of the entire SMR protein family are examined. Additionally, updated bioinformatic analysis of predicted and characterized SMR protein family members was also conducted. Based on SMR sequence alignments and phylogenetic analysis of current members, we propose that this small multidrug resistance transporter family should be expanded into three subclasses: (i) the small multidrug pumps (SMP), (ii) suppressor of groEL mutation proteins (SUG), and a third group (iii) paired small multidrug resistance proteins (PSMR). The roles of these three SMR subclasses are examined, and the well-characterized members, such as Escherichia coli EmrE and SugE, are described in terms of their function and structural organization.     

Comparison of the biochemical and molecular properties of myoglobins from three Biomphalaria species

Myoglobin is a globin with heme as prosthetic group whose main known biological role is to bind to O2 reversibly. On account of their large diversity, globins from mollusks have contributed to the study of this protein class. The cDNA of the myoglobins from Biomphalaria straminea and Biomphalaria tenagophila, which have a glutamine as distal residue (E7), were constructed and analyzed by bioinformatic tools. Native (not recombinant) myoglobins of these two Biomphalaria species were purified and their experimental molecular mass (about 16 kDa) and pI (about (8.0) were provided. Data analysis showed that these proteins are monomers with the signature for the classic myoglobin fold and they are blocked in amino terminus probably by an acetyl group. Values of the autoxidation rates showed that these myoglobins oxidized slowly. About the primary sequences of the myoglobins, they turned out to be satisfactory to group mollusks in phylogenetic class.     

Topoisomerase activity of the hyperthermophilic replication initiator protein Rep75

The plasmid pGT5 from the hyperthermophilic archaeon Pyrococcus abyssi replicates via the rolling circle mechanism. pGT5 encodes the replication initiator protein Rep75 that exhibits a nicking–closing (NC) activity in vitro on single-stranded oligonucleotides containing the pGT5 double-stranded origin (dso) sequence. Some mesophilic Rep proteins present site-specific DNA topoisomerase-like activity on a negatively supercoiled plasmid harbouring the dso. We report here that Rep75 also exhibits topoisomerase activity on a negatively supercoiled DNA substrate. This DNA topoisomerase-like activity is dependent on the amino acids involved in NC activity of Rep75. However, in contrast with mesophilic Rep proteins, Rep75 topoisomerase activity is not dso dependent. Moreover, although pGT5 is known to be relaxed in vivo, Rep75 was not able to act on a relaxed plasmid in vitro, whether or not it contained the dso.     

A gyrB-like gene from the hyperthermophilic bacterion Thermotoga maritima

We have cloned and sequenced two overlapping DNA fragments (3236 bp) containing a gene encoding the ATPase subunit of a type II DNA topoisomerase from the hyperthermophilic bacterion Thermotoga maritima (Tm Top2B). The deduced protein is composed of 636 aa with a calculated molecular mass of 72 415 Da. It shares significant similarities with the ATPase subunits of mesophilic bacterial DNA topoisomerases II, either DNA gyrase (GyrB) or DNA topoisomerase IV (ParE). Although the highest similarity scores are obtained with GyrB proteins (55% identity with Bacillus subtilis DNA gyrase), a detailed phylogenetic analysis of all known DNA topoisomerases II does not allow us to determine if Tm Top2B corresponds to a DNA gyrase or a DNA topoisomerase IV. This hyperthermophilic Top2B protein exhibits a larger amount of charged amino acids than its mesophilic homologues, a feature which could be important for its thermostability. No gyrA-like gene has been found near top2B. A gene coding for a transaminase B-like protein was found in the upstream region of top2B.     

Dynamics and unfolding pathways of a hyperthermophilic and a mesophilic rubredoxin.

Molecular dynamics simulations in solution are performed for a rubredoxin from the hyperthermophilic archaeon Pyrococcus furiosus (RdPf) and one from the mesophilic organism Desulfovibrio vulgaris (RdDv). The two proteins are simulated at four temperatures: 300 K, 373 K, 473 K (two sets), and 500 K; the various simulations extended from 200 ps to 1,020 ps. At room temperature, the two proteins are stable, remain close to the crystal structure, and exhibit similar dynamic behavior; the RMS residue fluctuations are slightly smaller in the hyperthermophilic protein. An analysis of the average energy contributions in the two proteins is made; the results suggest that the intraprotein energy stabilizes RdPf relative to RdDv. At 373 K, the mesophilic protein unfolds rapidly (it begins to unfold at 300 ps), whereas the hyperthermophilic does not unfold over the simulation of 600 ps. This is in accord with the expected stability of the two proteins. At 473 K, where both proteins are expected to be unstable, unfolding behavior is observed within 200 ps and the mesophilic protein unfolds faster than the hyperthermophilic one. At 500 K, both proteins unfold; the hyperthermophilic protein does so faster than the mesophilic protein. The unfolding behavior for the two proteins is found to be very similar. Although the exact order of events differs from one trajectory to another, both proteins unfold first by opening of the loop region to expose the hydrophobic core. This is followed by unzipping of the beta-sheet. The results obtained in the simulation are discussed in terms of the factors involved in flexibility and thermostability.     

Quantitative Mass Spectrometric Profiling of Cancer-cell Proteomes Derived From Liquid and Solid Tumors

In-depth analyses of cancer cell proteomes are needed to elucidate oncogenic pathomechanisms, as well as to identify potential drug targets and diagnostic biomarkers. However, methods for quantitative proteomic characterization of patient-derived tumors and in particular their cellular subpopulations are largely lacking. Here we describe an experimental set-up that allows quantitative analysis of proteomes of cancer cell subpopulations derived from either liquid or solid tumors. This is achieved by combining cellular enrichment strategies with quantitative Super-SILAC-based mass spectrometry followed by bioinformatic data analysis. To enrich specific cellular subsets, liquid tumors are first immunophenotyped by flow cytometry followed by FACS-sorting; for solid tumors, laser-capture microdissection is used to purify specific cellular subpopulations. In a second step, proteins are extracted from the purified cells and subsequently combined with a tumor-specific, SILAC-labeled spike-in standard that enables protein quantification. The resulting protein mixture is subjected to either gel electrophoresis or Filter Aided Sample Preparation (FASP) followed by tryptic digestion. Finally, tryptic peptides are analyzed using a hybrid quadrupole-orbitrap mass spectrometer, and the data obtained are processed with bioinformatic software suites including MaxQuant. By means of the workflow presented here, up to 8,000 proteins can be identified and quantified in patient-derived samples, and the resulting protein expression profiles can be compared among patients to identify diagnostic proteomic signatures or potential drug targets.     

Structure of PP4397 Reveals the Molecular Basis for Different c-di-GMP Binding Modes by Pilz Domain Proteins

Cyclic diguanylate (c-di-GMP) is a global regulator that modulates pathogen virulence and biofilm formation in bacteria. Although a bioinformatic study revealed that PilZ domain proteins are the long-sought c-di-GMP binding proteins, the mechanism by which c-di-GMP regulates them is uncertain. Pseudomonas putida PP4397 is one such protein that contains YcgR-N and PilZ domains and the apo-PP4397 structure was solved earlier by the Joint Center for Structural Genomics. We determined the crystal structure of holo-PP4397 and found that two intercalated c-di-GMPs fit into the junction of its YcgR-N and PilZ domains. Moreover, c-di-GMP binding induces PP4397 to undergo a dimer-to-monomer transition. Interestingly, another PilZ domain protein, VCA0042, binds to a single molecule of c-di-GMP, and both its apo and holo forms are dimeric. Mutational studies and the additional crystal structure of holo-VCA0042 (L135R) showed that the Arg122 residue of PP4397 is crucial for the recognition of two molecules of c-di-GMP. Thus, PilZ domain proteins exhibit different c-di-GMP binding stoichiometry and quaternary structure, and these differences are expected to play a role in generating diverse forms of c-di-GMP-mediated regulation.     

In vitro rapid evolution of fungal immunomodulatory proteins by DNA family shuffling

Fungal immunomodulatory proteins (FIPs) found in a wide variety of mushrooms hold significant therapeutic potential. Despite much research, the structural determinants for their immunomodulatory functions remain unknown. In this study, a DNA shuffling technique was used to create two shuffled FIP protein libraries: an intrageneric group containing products of shuffling between FIP-glu (FIP gene isolated from Ganoderma lucidum) and FIP-gsi (FIP gene isolated from Ganoderma sinense) genes and an intergeneric group containing the products of shuffling between FIP-glu, FIP-fve (FIP gene isolated from Flammulina velutipes), and FIP-vvo (FIP gene isolated from Volvariella volvacea) genes. The gene shuffling generated 426 and 412 recombinant clones, respectively. Using colony blot analysis, we selected clones that expressed relatively high levels of shuffled gene products recognized by specific polyclonal antibodies. We analyzed the DNA sequences of the selected shuffled genes, and testing of their protein products revealed that they maintained functional abilities to agglutinate blood cells and induce cytokine production by splenocytes from Kunming mice in vitro. Meanwhile, the relationships between protein structure and the hemagglutination activity and between the changed nucleotide sites and expression levels were explored by bioinformatic analysis. These combined analyses identified the nucleotide changes involved in regulating the expression levels and hemagglutination activities of the FIPs. Therefore, we were able to generate recombinant FIPs with improved biological activities and expression levels by using DNA shuffling, a powerful tool for the generation of novel therapeutic proteins and for their structural and functional studies.     

Identification of Human N-Myristoylated Proteins from Human Complementary DNA Resources by Cell-Free and Cellular Metabolic Labeling Analyses

To identify physiologically important human N-myristoylated proteins, 90 cDNA clones predicted to encode human N-myristoylated proteins were selected from a human cDNA resource (4,369 Kazusa ORFeome project human cDNA clones) by two bioinformatic N-myristoylation prediction systems, NMT-The MYR Predictor and Myristoylator. After database searches to exclude known human N-myristoylated proteins, 37 cDNA clones were selected as potential human N-myristoylated proteins. The susceptibility of these cDNA clones to protein N-myristoylation was first evaluated using fusion proteins in which the N-terminal ten amino acid residues were fused to an epitope-tagged model protein. Then, protein N-myristoylation of the gene products of full-length cDNAs was evaluated by metabolic labeling experiments both in an insect cell-free protein synthesis system and in transfected human cells. As a result, the products of 13 cDNA clones (FBXL7, PPM1B, SAMM50, PLEKHN, AIFM3, C22orf42, STK32A, FAM131C, DRICH1, MCC1, HID1, P2RX5, STK32B) were found to be human N-myristoylated proteins. Analysis of the role of protein N-myristoylation on the intracellular localization of SAMM50, a mitochondrial outer membrane protein, revealed that protein N-myristoylation was required for proper targeting of SAMM50 to mitochondria. Thus, the strategy used in this study is useful for the identification of physiologically important human N-myristoylated proteins from human cDNA resources.     

Thermal stability of CopA, a polytopic membrane protein from the hyperthermophile Archaeoglobus fulgidus

Despite recent progress in understanding membrane protein folding, little is known about the mechanisms stabilizing these proteins. Here we characterize the kinetic thermal stability of CopA, a thermophilic P_IB-type Cu⁺-ATPase from Archaeoglobus fulgidus. When heterologously expressed in Escherichia coli, purified and reconstituted in mixed micelles, CopA retained thermophilic characteristics with maximum activity at 75 °C. Incubation of CopA in the absence of substrates at temperatures in the 66-85 °C range led to an irreversible exponential decrease in enzyme activity suggesting a two-state process involving fully-active and inactive molecules. Although CopA inactivated much slower than mesophilic proteins, the activation energy was similar to that observed for mesophilic P-type ATPases. The inactivation process was found to be associated with the irreversible partial unfolding of the polypeptide chain, as assessed by Trp fluorescence, Phe UV spectroscopy, far UV circular dichroism, and 1-aniline-8-naphtalenesulfonate binding. However, the inactive thermally denatured protein still conserves large hydrophobic regions and considerable secondary structure.