Comparative model-building of the mammalian serine proteases

Proteins have been classified into families based upon sequence homology. An accurate, systematic comparative model-building procedure for a homologous family of proteins would be very valuable scientifically. This paper presents such a procedure and applies it to the mammalian serine proteases, which are ubiquitous and involved in many important biological functions. Eleven proteins of this family are considered here, including a variety of blood serum, intestinal and pancreatic proteins as well as a closely related bacterial enzyme.The modeling method capitalizes upon the availability of three experimentally determined structures for mammalian serine proteases. These structures show that the molecule is divided into structurally conserved regions, which contain the strong sequence homology, and structurally variable regions, which include all the additions and deletions. We show that by applying this structural distinction to new sequences, erroneous alignments of the sequences are greatly minimized.For each aligned new sequence, the structurally conserved regions can be constructed from any of the known structures. In examining the variable regions, we have found that a variable region that has the same length and residue character in two different known structures usually has the same conformation in both. Thus, when the eight structurally unknown proteins are modeled, most of the variable regions can be constructed directly from the known structures. A minority of the variable regions require more sophisticated analysis to evaluate the relative merits of a small number of possible conformations. Only a very few are so different that modeling by homology is entirely ruled out. We demonstrate, therefore, that by this modeling procedure, the maximum of each of these mammalian serine proteases is constructed directly from the experimentally determined structures and the necessity to build from intuition or from energy considerations is greatly reduced.     

Crystallographic protein model-building on the web

X-ray crystallography is the most widely used method to determine the 3D structure of protein molecules. One of the most difficult steps in protein crystallography is model-building, which consists of constructing a backbone and then amino acid side chains into an electron density map. Interpretation of electron density maps represents a major bottleneck in protein structure determination pipelines, and thus, automated techniques to interpret maps can greatly improve the throughput. We have developed WebTex, a simple and yet powerful web interface to TEXTAL, a program that automates this process of fitting atoms into electron density maps. TEXTAL can also be downloaded for local installation. Availability: Web interface, downloadable binaries and documentation at http://textal.tamu.edu     

Computer-aided model-building strategies for protein design

Model-building strategies for protein modification and design are developed. These strategies emphasize simple geometric aspects of protein structure, use local coordinate systems defined at particular residues, and systematically consider a large number of alternative sequences and conformations. We have written a computer program, PROTEUS, to implement these search methods. PROTEUS has been used to find positions where disulfide bonds could be added to the N-terminal domain of the lambda repressor and to predict how a loop on the surface of repressor could be shortened.     

'Pals'. A medical student public service program.

We designed a public service and educational program to aid children and families coping with chronic illness and to augment medical student education. Medical students developed relationships with chronically ill children and families based on the Big Brother-Big Sister program model. In addition, students attended bimonthly seminars on childhood chronic illness and family dynamics. Medical students learned about the psychosocial aspects of illness through these relationships and reported that the program contributed to their sense of worth as caregivers. By fostering students'' innate altruism, medical schools may succeed in cultivating caring and humanism in their student physicians. We propose a model that encourages medical students to relate personally with patients and their families. A program such as this has the potential to nurture compassion in medical students, contribute to medical education, and provide support to patients and families.     

Molecular model-building of amylin and alpha-calcitonin gene-related polypeptide hormones using a combination of knowledge sources

Amylin is the major component of the amyloid found in the pancreases of noninsulin-dependent diabetics (type 2 diabetes). It is a 37 amino acid polypeptide and has been shown to have 46% sequence identity with the neuropeptide alpha-calcitonin gene-related peptide (alpha-CGRP). Both amylin and alpha-CGRP are known to be potent inhibitors of glycogen synthesis in stripped rat soleus muscle. Secondary structure prediction and tertiary structure model-building show the two polypeptides to have an alpha-helix/beta-strand motif similar to that observed in the insulin B-chain. The results have been supported by CD spectroscopy, although there is no sequence similarity between insulin and amylin/alpha-CGRP. Aggregation states have been predicted based on the dimeric and hexameric arrangements seen in porcine insulin. Rat and hamster amylin have a changed sequence motif in the beta-strand which results in lack of amyloid formation and type 2 diabetes. This, we propose, is caused by disruption of hydrogen bonding which prevents the formation of the dimer.     

显微系统化与大学生科研培训计划中科研思维的培养

显微系统化是指将各种显微镜融合并统一利用,联合操作,所用程序一体化、规范化、统一化和数字化。显微系统化通过系统化管理,在大学生科研培训计划（Student Research Training Program,SRTP）中实施。实行显微系统化,可突破显微镜本身时间和空间上的局限性,能够进一步完善科学、高效的科研培训计划模式,对从感性方面培养大学生理解科研的概念和建立科研思维的能力、提高大学生科研素养具有重要作用。     

X-ray and model-building studies on the specificity of the active site of proteinase K

Proteinase K, the extracellular serine endopeptidase (E.C. 3.4.21.14) from the fungus Tritirachium album limber, is homologous to the bacterial subtilisin proteases. The binding geometry of the synthetic inhibitor carbobenzoxy-Ala-Phechloromethyl Ketone to the active site of proteinase K was the first determined from a Fourier synthesis based on synchrotron X-ray diffraction data between 1.8 Å and 5.0 Å resolution. The protein inhibitor complexes was refined by restrained least-squares minimization with the data between 10.0 and 1.8 Å. The final R factor was 19.1% and the model contained 2,018 protein atoms, 28 inhibitors atoms, 125 water molecules, and two Ca²⁺ ions. The peptides portion of the inhibitor is bound to the active center of proteinase K by means of a three-stranded antiparallel pleated sheet, with the side chain of the phenylalanine located in the P1 site. Model building studies, with lysine replacing phenylalanine in the inhibitor, explain the relatively unspecific catalytic activity of the enzyme.     

Darwin as a student of behavior

In The Expression of the Emotions, Charles Darwin documents evolutionary continuity between animals and humans, emphasizing the universality of expressions in man. Most of the book addresses human behavior, and its influence on the study of animal behavior has been weak. The issue of natural selection is remarkably absent from this book, which relies on the inheritance of acquired characters rather than on a genuine Darwinian logic. Yet Konrad Lorenz considered Darwin to be a forerunner of behavioral biology. The reason was to be found in The Descent of Man and chapter VIII of The Origin of Species, where Darwin provides an explanation of behavior through selection, stating that the same mechanisms explaining morphological changes also account for gradual improvements in instincts. He assessed the accuracy of his evolutionary theory by directly studying animal behavior, hence laying the foundations of behavioral research for the next century.     

Electrostatic calculations and model-building suggest that DNA bound to CAP is sharply bent

Two observations suggest that DNA, upon binding to E. coli catabolite gene activator protein (CAP), is sharply bent by a total angle of at least 100-150 degrees: (1) The electrostatic potential field of CAP shows regions of positive potential that form a ramp on 3 sides of the protein. (2) The DNA binding site size as determined by DNA ethylation interference with binding, (Majors: "Control of the E. coli Lac Operon at the Molecular Level." Ph.D. Thesis, Harvard University, Cambridge, 1977) and by relative affinities of DNA fragments of various lengths (Liu-Johnson et al.: Cell 47:995-1005, 1986) requires severe bending of the DNA to maintain its favorable electrostatic contact with the protein.     

X-ray fiber diffraction and model-building study of polyguanylic acid and polyinosinic acid.

X-ray diffraction patterns of fibers of polyriboguanylic acid and polyriboinosinic acid are shown to be virtually identical. These diffraction patterns are consistent only with three or four-stranded models. Model-building studies on a computer-assisted interactive display system favor the four-stranded model. In addition, the greater thermal stability of poly(rG) relative to poly(rI) can be accounted for by a four-stranded model in which there are two hydrogen bonds per base for poly(rG) versus one for poly(rI).