Importance of environment in determining secondary structure in proteins.
Importance of environment in determining secondary structure in proteins.
复制标题
环境在确定蛋白质二级结构中的重要性。
DOI：
10.1021/bi00174a019
复制
发表时间：
1994
期刊：
影响因子：
2.9
通讯作者：
JohnsonJr,WC
中科院分区：
文献类型：
--
作者：
Waterhous,DV;JohnsonJr,WC
修订稿于 1993 年 12 月 1 日收到*摘要：我们在此报告本体溶剂环境对几种肽二级结构的影响。在之前的工作中，通过圆二色性 (CD) 光谱法，根据氨基酸偏好预测为螺旋状但在其蛋白质中发现为 3 链的可疑肽序列在醇溶剂中显示为螺旋，在非胶束十二烷基硫酸钠 (SDS) 中为 3 链 [Zhong, L., & Johnson, W. C., Jr.(1992) Proc.国家。阿卡德。科学。美国 89, 4462-4465]。在这里，我们展示了预测为 3 链但被发现为螺旋的模棱两可的序列遵循相同的模式；它们在醇溶剂中为螺旋状，在非胶束 SDS 中为 3 链。此外，我们研究了仅具有强螺旋倾向的控制序列和仅具有强/3链倾向的控制序列。这两个表现良好的序列都遵循与模棱两可的序列相同的模式。与其他肽类似，在所有溶剂中异常稳定的 Y (EAAAK^A 都是反螺旋，但 CD 谱分析表明，在将溶剂从三氟乙醇 (TFE) 更改为 SDS 时，螺旋会随着 /3 链和其他结构的增加而丢失。我们发现，溶剂是体外确定氨基酸序列二级结构的一个非常重要的因素，并且可以覆盖由于序列而产生二级结构的倾向。这意味着二级结构所见的微溶剂是由于从蛋白质的三级结构到氨基酸的非局部相互作用（我们称之为环境）可能是决定肽二级结构的重要因素，因此应该考虑正确预测蛋白质中氨基酸序列的二级结构。蛋白质的序列清楚地决定了其天然结构，而局部相互作用形式的序列已成为预测二级结构的流行因素，许多关注于从一级序列先验预测蛋白质二级结构的研究工作都是从统计数据中发展而来的。已知蛋白质结构（Chou & Fasman，1978；Burgess 等，1974；Lim，1974；Gamier 等，1978）或同源序列比较（Pongor & Szaley，1985；Sweet，1986；Nishikawa & Ooi，1986；Levin 等，1986；Zvelebil 等，这些方法相当成功，这刺激了人们继续研究确定哪些因素控制氨基酸序列折叠成球状蛋白质。这种努力导致了一些精心设计的模型蛋白质的开发，这些蛋白质折叠成预测的二级蛋白质（Lau 等人，1984；Eisenberg 等人，1986；Ho 和 DeGrado，1987；Marqusee 和 Baldwin，1987；Richardson 和 Richardson 等人。 Richardson, 1987; Lyu et al., 1989)、三级结构 (Ho & DeGrado, 1987) 和四级结构 (Hill et al., 1990) 的成功让许多研究人员相信我们正在解决蛋白质折叠问题，因此我们确实能够破解遗传密码的后半部分。这比单纯的百分比暗示的要好得多，因为许多错误都位于正确预测的二级结构的末端。然而，一些序列是根据氨基酸偏好在一个二级结构中预测的，但却被发现。
Revised Manuscript Received December 1, 1993* abstract: We report here the effect of bulk solvent environment on the secondary structure of several peptides. In previous work, equivocal peptide sequences that are predicted to be-helical from amino acid preference but are found to be/3-strand in their proteins were shown to be-helical in alcoholsolvents and/3-strand in nonmicellar sodium dodecyl sulfate (SDS) by circular dichroism (CD) spectroscopy [Zhong, L., & Johnson, W. C., Jr.(1992) Proc. Natl. Acad. Sci. USA 89, 4462-4465]. Here we show that equivocal sequences that are predicted to be/3-strand but are found to be-helical follow the same pattern; they are-helical in alcoholsolvents and/3-strand in nonmicellar SDS. Furthermore, we investigated a control sequence with only a strong-helical propensity and a control sequence with only a strong/3-strand propensity. Both of these well-behaved sequences followed the same pattern as the equivocal sequences. The exceptionally stable Y (EAAAK^ A is an-helix in all solvents, but analyses of the CD spectra indicate the loss of helix with an increase in/3-strand and other structures on changing solvent from trifluoroethanol (TFE) to SDS, similar to the other peptides. We find that solvent is a very important factor in determining the secondary structure of an amino acid sequence in vitro and can override the propensity for a secondary structure due to sequence. This implies that the microsolvent seen by a secondary structure due to nonlocal interactions of amino acids from the tertiary structure of a protein, which we call environment, may be an important factor in determining the secondary structure of peptides and therefore should be considered to correctly predict thesecondary structure of an amino acid sequence in proteins.The sequence of a protein clearly determines its native structure, andsequence, in the formof local interactions, has been a popular factor for predicting secondary structure. Much of the research effort that has focused on a priori prediction of protein secondary structures from their primary sequences has been developed from the statistics of known protein structures (Chou & Fasman, 1978; Burgess et al., 1974; Lim, 1974; Gamier et al., 1978) or comparison of homologous sequences (Pongor & Szaley, 1985; Sweet, 1986; Nishikawa & Ooi, 1986; Levin et al., 1986; Zvelebil et al., 1987). These methods are rather successful, and this has stimulated the continued research effort into determining what factors control the folding of a sequence of amino acids into a globular protein. This effort has resulted in the development of some well-designed model proteins, which fold into predicted secondary (Lau et al., 1984; Eisenberg et al., 1986; Ho & DeGrado, 1987; Marqusee & Baldwin, 1987; Richardson & Richardson, 1987; Lyu et al., 1989), tertiary (Ho & DeGrado, 1987), and quarternary (Hill et al., 1990) structures de novo. Such successes make many researchers confidentthat we are solving the protein folding problem, and therefore that we will indeed be able to crack the second half of the genetic code. Current a priori methods are about 70% successful at predicting secondarystructure from primary sequence; this is much better than mere percentages imply, because many of the errors are at the ends of correctly predicted secondary structures. However, some sequences are predicted in one secondary structure from amino acid preferences but are found