Bioinformatics Advance Access originally published online on June 10, 2008
Bioinformatics 2008 24(16):1819-1820; doi:10.1093/bioinformatics/btn255
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
MINS2: Revisiting the molecular code for transmembrane-helix recognition by the Sec61 translocon
Center for Bioinformatics, Saarland University, Germany
*To whom correspondence should be addressed.
 |
ABSTRACT |
|---|
 TOP ABSTRACT 1 INTRODUCTION2 DEVELOPMENT AND BENCHMARK...3 APPLICATION OF MINS2ACKNOWLEDGEMENTSREFERENCES |
|---|
Summary: To be fully functional, membrane proteins should not only fold, but also get inserted into the membrane, which is mediated by the Sec61 translocon. Recent experimental studies have attempted to elucidate how the Sec61 translocon accomplishes this delicate task by measuring the translocon-mediated membrane insertion free energies of 357 systematically designed peptides. On the basis of this data set, we have developed MINS2, a novel sequence-based computational method for predicting the membrane insertion free energies of protein sequences. A benchmark analysis of MINS2 shows that MINS2 significantly outperforms previously proposed methods. Importantly, the application of MINS2 to known membrane protein structures shows that a better prediction of membrane insertion free energies does not lead to a better prediction of transmembrane segments of polytopic membrane proteins.
Availability: A web server for MINS2 is publicly available at http://service.bioinformatik.uni-saarland.de/mins.
Contact: volkhard.helms{at}bioinformatik.uni-saarland.de
Supplementary information: Supplementary data are available at Bioinformatics online.
|
1 INTRODUCTION |
|---|
|
TOPABSTRACT1 INTRODUCTION2 DEVELOPMENT AND BENCHMARK...3 APPLICATION OF MINS2ACKNOWLEDGEMENTSREFERENCES |
|---|
Membrane proteins play a crucial role in various cellular processes. Unlike water-soluble proteins, membrane proteins should not only fold, but also get integrated into the membrane to be fully functional. The Sec61 translocon, a protein complex residing in the endoplasmic reticulum membrane, mediates this membrane insertion process (Hessa et al., 2005; van den Berg et al., 2004). More specifically, the Sec61 translocon discriminates between transmembrane (TM) and non-TM segments, inserting TM segments into the membrane while letting non-TM segments pass by. Recent experimental studies have attempted to elucidate how the Sec61 translocon accomplishes this delicate task by measuring the translocon-mediated membrane insertion free energies of 357 systematically designed peptides (Hessa et al., 2007). Using this important data set, we have developed MINS2, a highly accurate sequence-based computational method for predicting the membrane insertion free energies of protein sequences.
Interestingly, the recent experimental studies by von Heijne and his coworkers have been controversial, some questioning the relevance of their results to the Sec61 translocon's activity in the cell (Dorairaj and Allen, 2007; MacCallum et al., 2007; Shental-Bechor et al., 2006). Thus, it is unclear how relevant the data on the 357 peptides (hereafter denoted as the TH data) are to the in vivo biogenesis of membrane proteins by the Sec61 translocon. As shown subsequently, the application of MINS2 to known membrane protein structures unravels hitherto unappreciated aspects about the TH data, which may contribute to clarifying the controversy over them.
|
2 DEVELOPMENT AND BENCHMARK ANALYSIS OF MINS2 |
|---|
|
TOPABSTRACT1 INTRODUCTION2 DEVELOPMENT AND BENCHMARK...3 APPLICATION OF MINS2ACKNOWLEDGEMENTSREFERENCES |
|---|
The full details of how MINS2 was developed are available in the supplementary information. For space reasons, only a brief description is given here. The development of MINS2 was based on the TH data, where all peptides are 19 residues long. One of the simplest models for the TH data would thus be a linear model having 380 (=19 x 20) terms along with an intercept. In fact, this formed the basis of the two previous approaches, MINS1 (Park and Helms, 2008) and the prediction method from von Heijne (Hessa et al., 2007) [hereafter denoted as the TH method]. First, prompted by previous free energy profiles from known structures (Senes et al., 2007; Ulmschneider et al., 2005) as well as from the TH data (Hessa et al., 2007), the linear model was simplified to have 101 terms. Then, inspired by the surprising power of the polynomial kernel of degree 2 in some applications (Hastie et al., 2001), we identified beneficial non-self cross terms. This type of cross-interactions has in fact been shown to play an important role in membrane insertion of TM segments, e.g. the motif of an Asp residue and a Lys residue spaced apart three residues (Chin and von Heijne, 2000). The resultant model (MINS2) is a simple linear regression that takes into account 340 terms and an intercept. MINS2 is neither a kernalized linear regression nor a support vector regression (see Supplementary information for the details). MINS2 was tested for the TH data using a leave-one-out scheme. As shown in Table 1, the correlation coefficient between MINS2-predicted and experimentally measured membrane insertion free energies is 0.94. The mean unsigned error between them is 0.20 kcal/mol. MINS2 significantly outperforms the two previously proposed sequence-based prediction methods [P<1.72 x 10–6 for the TH method and P<2.20x10–16 for MINS1; the prediction results of the TH method and MINS1 were taken from the respective publications (Hessa et al., 2007; Park and Helms, 2008)].
|
|
3 APPLICATION OF MINS2 |
|---|
|
TOPABSTRACT1 INTRODUCTION2 DEVELOPMENT AND BENCHMARK...3 APPLICATION OF MINS2ACKNOWLEDGEMENTSREFERENCES |
|---|
In addition to allowing us to accurately predict the translocon-mediated membrane insertion free energies of protein sequences, we have tested whether MINS2, along with the two previous methods, can be used to help address the controversial issue of how relevant the TH data are to the Sec61 translocon's activity in the cell for generating membrane proteins. Our reasoning is as follows. If the TH data convey information relevant to the Sec61 translocon's activity, it is expected that the more accurately a prediction method extracts the information in the TH data, the more accurately it should be able to predict TM segments in protein sequences. This is so because the discrimination between TM and non-TM segments is the fundamental task of the Sec61 translocon in generating membrane proteins. Hence, the three methods were tested for their ability to predict TM segments as described in the Supplementary information. As shown in Table 1, the three methods performed nearly equally for bitopic membrane proteins. For polytopic membrane proteins, however, the more accurately a prediction method captures the information in the TH data, the more poorly it performs in predicting TM segments.
These results suggest the following point. Since the TH data were derived from the insertion behavior of single TM segments, one may expect them to work well for bitopic membrane proteins, and this is what is observed in Table 1. However, the membrane insertion behavior of individual TM segments may not fully account for the membrane insertion of polytopic membrane proteins. A simple calculation hints at this point. For a membrane protein with seven TM segments, if one assumes a membrane insertion probability of 0.7 for individual TM segments, then the overall biogenesis efficiency is less than 0.09, which is unacceptably low. In fact, the typical membrane insertion probability of individual TM segments of polytopic membrane proteins was suggested to be less than 0.7 (Hessa et al., 2007; Park and Helms, 2008). In this regard, it is also worth noting that the TM segments of bitopic membrane proteins have been suggested to tend to possess more favorable membrane insertion free energies than those of polytopic membrane proteins (Arkin and Brunger, 1998; Hessa et al., 2007; Park and Helms, 2008). Overall, this line of reasoning suggests that the TH data are not expected to work equally well for polytopic membrane proteins.
Of course, there are more ways of interpreting the results reported here, and further experimental and computational studies are needed to fully resolve the controversial issue of how relevant the TH data are to the Sec61 translocon's real activity. The current study provides a stepping stone towards this long-term goal by revealing hitherto unappreciated aspects of the TH data with regard to detecting TM segments of polytopic versus bitopic membrane proteins.
We hope that experimental and theoretical researchers working on topological analysis of helical membrane proteins will find the MINS2 web server useful for their work.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACT1 INTRODUCTION2 DEVELOPMENT AND BENCHMARK...3 APPLICATION OF MINS2ACKNOWLEDGEMENTSREFERENCES |
|---|
We thank crystallographers of membrane proteins because the current work would have been impossible without their work.
Conflict of Interest: none declared.
|
FOOTNOTES |
|---|
Associate Editor: Anna Tramontano
Received on April 25, 2008; revised on June 2, 2008; accepted on June 2, 2008
|
REFERENCES |
|---|
|
TOPABSTRACT1 INTRODUCTION2 DEVELOPMENT AND BENCHMARK...3 APPLICATION OF MINS2ACKNOWLEDGEMENTSREFERENCES |
|---|
Arkin IT, Brunger AT. Statistical analysis of predicted transmembrane alpha-helices. Biochim. Biophys. Acta (1998) 1429:113–128.[CrossRef][Medline]
Chin CN, von Heijne G. Charge pair interactions in a model transmembrane helix in the ER membrane. J. Mol. Biol (2000) 303:1–5.[CrossRef][Web of Science][Medline]
Dorairaj S, Allen TW. On the thermodynamic stability of a charged arginine side chain in a transmembrane helix. Proc. Natl Acad. Sci. USA (2007) 104:4943–4948.
Hastie T, et al. The Elements of Statistical Learning. (2001) New York: Springer.
Hessa T, et al. Recognition of transmembrane helices by the endoplasmic reticulum translocon. Nature (2005) 433:377–381.[CrossRef][Web of Science][Medline]
Hessa T, et al. Molecular code for transmembrane-helix recognition by the Sec61 translocon. Nature (2007) 450:1026–1030.[CrossRef][Web of Science][Medline]
MacCallum JL, et al. Partitioning of amino acid side chains into lipid bilayers: results from computer simulations and comparison to experiment. J. Gen. Physiol (2007) 129:371–377.
Park Y, Helms V. Prediction of the translocon-mediated membrane insertion free energies of protein sequences. Bioinformatics (2008) 24:1271–1277.
Senes A, et al. E(z), a depth-dependent potential for assessing the energies of insertion of amino acid side-chains into membranes: derivation and applications to determining the orientation of transmembrane and interfacial helices. J. Mol. Biol (2007) 366:436–448.[CrossRef][Web of Science][Medline]
Shental-Bechor D, et al. Has the code for protein translocation been broken? Trends Biochem. Sci (2006) 31:192–196.[CrossRef][Web of Science][Medline]
Ulmschneider MB, et al. Properties of integral membrane protein structures: derivation of an implicit membrane potential. Proteins (2005) 59:252–265.[CrossRef][Web of Science][Medline]
van den Berg B, et al. X-ray structure of a protein-conducting channel. Nature (2004) 427:36–44.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  B. C. S. Cross and S. High Dissecting the physiological role of selective transmembrane-segment retention at the ER translocon J. Cell Sci., June 1, 2009; 122(11): 1768 - 1777. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||