Why proline not in alpha helix

Sancho, Org. To request permission to reproduce material from this article, please go to the Copyright Clearance Center request page. If you are an author contributing to an RSC publication, you do not need to request permission provided correct acknowledgement is given.

If you are the author of this article, you do not need to request permission to reproduce figures and diagrams provided correct acknowledgement is given. Read more about how to correctly acknowledge RSC content. Fetching data from CrossRef.

This may take some time to load. Loading related content. Jump to main content. Jump to site search. You do not have JavaScript enabled. Please enable JavaScript to access the full features of the site or access our non-JavaScript page.

Issue 4, The angles of each of the helices in a family can then be compared at every site in the alignment. Only one site of maximum disruption was used to classify a family.

The site in the helix with the highest mean was chosen as the most disrupted site. Only sites in the alignment of the angle data which had at least five recorded angles after smoothing were considered.

To determine the variation of angles at the most disrupted site, standard deviation was calculated. The standard deviation was compared to the mean error of angles at the most disrupted site in order to classify families as conserved or not. The flowchart in Fig 1 shows the method of classification for helix families. A package containing all data for the homologous helix pairs and families is available in S1 Data.

In this paper we analyse whether helix kinks are conserved between homologs by comparing their angles. The first step is to calculate a confidence interval on the helix kink angles measured.

We used our novel method within Kink Finder [ 1 ] to calculate the confidence interval for every angle see Methods. Fig 2A shows two helices which have a difference in angle of In Fig 2B , two helices are shown that have angles which differ by only Thus we consider these helices to have significantly different kink angles.

PDB code, chain identifier and residue numbers are given for each helix. The black residues are at the most disrupted site see Methods in each helix pair. Non-redundant sets of 18, soluble and membrane protein chains with at least one helix of 12 or more residues were collected.

From these, 4, aligned pairs of homologous membrane helices MemPairs and , aligned homologous soluble helix pairs SolPairs were extracted. Kink Finder was used to measure the angles and the error on these angles for all residues in every helix. The number of helix pairs in each of these classes is shown in Table 1. As expected, kinks are more common in the membrane protein set, probably because membrane helices are longer and longer helices are more frequently kinked [ 1 ].

We tested whether the four classes are different in terms of proline occurrence, as proline is commonly associated with kinks. Proline was identified as present if it was at the position of the largest angle or in the four following residues. Table 1 shows the presence of proline in the aligned helix pairs, broken down by pair type for a detailed breakdown see Table A in S1 Document.

For both membrane and soluble proteins, when proline is present in both helices, most pairs are Conserved Kinked. When proline is not present in either helix at the most disrupted site, Conserved Straight is most common. In the case where proline is present in a kinked helix but not conserved in its homologs, it has been suggested that the kink will be conserved throughout the family.

However, in our data, if a helix has a proline and is kinked and its partner helix does not have a proline, the pair may be Conserved Kinked or Not Conserved. Not Conserved is nearly as common as Conserved Kinked when P- and -P are combined for membrane helix pairs, and more common in soluble helix pairs, indicating that loss of proline is often accompanied by a significant reduction in kink angle.

As we are considering the angle difference between homologous helices, we wanted to investigate how it relates to the sequence identity between helices Fig 3A , the sequence identity between neighbouring residues Fig 3B , calculated as described in Methods , and the sequence identity between the complete chains Fig 3C , in all cases ignoring gaps.

As expected, there is a trend for larger angle differences to be associated with lower sequence identity. Conversely, kinks are well conserved when sequence is conserved, whether on a local or global scale. Data from the redundant membrane protein set is shown so that the full range of sequence identity can be seen. As homologous helix pairs showed large numbers of unconserved kinks, we built families of homologous helices in order to investigate kink conservation patterns across larger samples of related helices.

Families contained at least five members, and Table 2 gives the number of families of various sizes. Angles were measured for all members in a family and smoothed as described in the Methods. Each family was classified in an analogous way to the pair classification Fig 1. Fig 4 shows an example of one family from each class.

In soluble proteins, Conserved Kinked families are rare 72 compared to Not Conserved families , but in membrane proteins they are equally common 12 of each.

The grey box indicates the site of maximum disruption used to classify the helix see Methods. The relationship between prolines and kinks in our homologous helix families gives similar patterns to those seen for pairs Fig L in S1 Document. If proline is found at the most disrupted site or the four following residues in every member of a family, it is a good indicator of kink conservation, as it was for helix pairs.

Of the 10 membrane and soluble families where proline is fully conserved, 2 are classified as not conserved. For the membrane and soluble helix families where proline is present in some but not all members, families are more frequently Not Conserved than Conserved Kinked Thus, once again proline conservation does not equal kink conservation in every case, and proline loss may or may not equal kink loss.

In an analogous way to that used for pairs, we analysed the relationship between kink angle conservation and sequence conservation. Helix sequence identity was calculated in the same way using just the consensus helix positions. The relationship between family angle variation and sequence conservation Fig J in S1 Document is similar to that seen for pairs Fig 3.

There are very few data points at the higher end of the sequence identity range, even when combining all soluble and membrane data, as the family detection method led to most families containing at least some distant homologs. The partial correlation coefficient for helix sequence identity, given chain sequence identity as a controlling variable, is This reinforces the suggestion found with the pairs that local and global sequence changes are both associated with kink angle changes.

The most disrupted site in each helix see Methods is shown in grey, and the standard deviation and classification of each helix is given. The most disrupted site in each helix see Methods is shown in grey. TMH 2 has a kink at residue 2x55 which is present in almost all members, but shows a wide range of angles and is therefore classified as Not Conserved. TMHs 4 and 5 have a kink in most members at the most disrupted site, but like TMH 2, these kinks take a wide range of angles.

TMH 6 and 7 are the only helices which have a kink classified as conserved. They also have the largest kink angles Fig 6a. There also seems to be a not conserved kink present in a small number of members around residue 6x34, near the N-terminal end of the helix.

Colouring is by a mean and c standard deviation of angles at each site in the GPCR family, on a spectrum from the lowest values in blue to the highest in red. Grey residues have no angles measured as they are loop regions or within 6 residues of the end of the consensus helix the minimum for a cylinder fit. Fig 6 shows the average size of angles measured at each site and their conservation across the family, presented on a structure of rhodopsin.

In the GPCR set, multiple structures were available for some receptors. This made it possible to observe variation in angles for an individual receptor.

We could also compare the distribution of angles for one receptor to the distributions of others. The kink at 1x43 is in a functionally significant region and present in just a few GPCRs [ 31 ]. Three of these GPCRs have proline near the kink, but there is no obvious sequence causing the other five to be kinked. B Angle distribution at position 3x28 in the human adenosine A 2A receptor. Fig M in S1 Document displays the errors for the angle data from both histograms.

The full set of GPCR structures was also separated based on the type of ligand bound, but there was no difference between agonist, antagonist and inverse agonist structures overall. However, there were four receptors for which at least ten structures were available in the GPCRDB, so comparisons could be made between the different activation states of these individual receptors.

This site has the highest standard deviation of any angle in any of the seven GPCR helices coloured red in Fig 6c. In the case of the A 2A receptor, the angle change is from straight in the agonist-bound structures to kinked in the antagonist-bound structures Fig 7B. This suggests that the change in angle is involved in the conformational change which occurs on activation of the receptor. The kink location in the helix is at the binding site for the natural ligand, therefore in this case its flexibility seems to be particularly important for the function of the receptor.

This change of helix shape has previously been described qualitatively [ 32 ]. Our method also allows us to identify correlations between the kink angles in different helices. These could suggest concerted motion or interaction between the helices, where the kinking of one helix affects the structure of the other. The sites of these kinks are slightly separated in the structure with TMH 4 between them Fig 6c.

There is a weaker correlation between the angle at these sites and the TMH 4 kink: 0. These correlations suggest that change of conformation in one helix can influence the conformation of another. Using Kink Finder, we have developed a new method of error estimation which makes it possible to compare helices and state whether their angles are different.

The method, which is based on the quality of fit to the helix either side of the kink, is to our knowledge the first that is able to obtain a confidence interval on a measured kink angle. Kink Finder uses a method of fitting cylinders to either side of the kink, which has the advantage of reliably finding changes of direction even in the presence of non-canonical hydrogen bonding regions.

This method requires a helix of 12 residues or more, but it is known that longer helices are more frequently kinked [ 1 ]. This results in a single metric, the kink angle, which allowed us to understand the error distribution and state the statistical significance of a difference in angles.

In this work, for an overview of kink conservation, we chose to classify each helix pair or family using one most disrupted site. This avoids biasing the results, however it represents a simplification of the more detailed information shown in Fig 5.

As we show for GPCRs, analysing all positions in a helix of interest reveals more about the system. In order to facilitate such analysis, this data is available for all of our helix families S1 Data. Using the error estimation method, we have shown that kinks are not always conserved across structural homologs, i. The different conformations seen in a pair or family can be explained by two possibilities, or a combination of both: The differences in sequence between two related proteins causes them to adopt different conformations.

There is conformational flexibility at the kink, which could be important for function. An investigation of the sequence dependence of the changes in angle provides evidence in support of the first option.

The exact sequence drivers for kinks remain elusive. Changes in conformation would be important to understand when modelling a homolog with no proline where the template has a proline kink, or vice versa.

More generally, changes in angle are associated with changes in sequence. We found that this relationship between kink conservation and sequence conservation is similar, whether considering only the residues in the homologous helices themselves, the residues in spatial proximity to the homologous helices, or the complete chains of the homologous proteins. Current kink modelling approaches assume that global effects are important for kink formation, especially for predicting the size of kinks [ 33 ].

Amino acids with a side chain whose movement is largely restricted in an alpha helix branched at beta carbon like threonine or valine are disfavored, i. Glycine, with its many possible main chain conformations, is also rarely found in helices. Knowing how likely an amino acid is to occur in an alpha helix the so-called helix propensities , it is possible to predict where helices occur in a protein sequence.

Proline is considered a helix breaker because its main chain nitrogen is not available for hydrogen bonding. Here is an example of a at the position of a. Prolines are often found near the beginning or end of an alpha helix, as in this example of this is an ultra high resolution structure where hydrogen atoms - white - are resolved and some atoms are shown in multiple positions.

The beginnings and ends of helices are called N-caps and C-caps, respectively, and they have interesting sequence and structural patterns involving main chain or side chain hydrogen bonding.

The first two protein structure to be determined, myoglobin and hemoglobin , consists mainly of alpha helices. Researchers were surprised to see how random the orientation of helices seemed to be. Other all alpha-helical proteins show bundles of nearly parallel or antiparallel helices e. In structures that have beta sheets and alpha helices, one common fold is a single beta sheet that is sandwiched by layers of alpha helices on either side for example Carboxypeptidase A.

When an alpha helix runs along the surface of the protein, one side of it will show polar side chains solvent accessible while the other side will show non-polar side chains part of the hydrophobic core. The alpha helix fits nicely into the major groove of DNA.

Many common DNA-binding motifs, such as the helix-turn-helix e. FIS protein or the zinc finger motif e. A common fold found in transmembrane proteins are alpha-helical bundles running from one side to the other side of the membrane. An alpha helix of 19 amino acids with a length of about 30 angstroms has the right size to cross the double-layer of a typical membrane. If the helix runs at an angle instead of perfectly perpendicular to the membrane, it has to be a bit longer.