Computer aided modeling of a fructose repressor

Transcription

1 Computer aided modeling of a fructose repressor Miia Helanto and Kristiina Kiviharju Abstract The bioconversion of fructose to mannitol is a commercially interesting bioprocess. Fructose can, however be utilized also otherwise, and this utilization is partly controlled by a fructose repressor (frur). In this study, we calculated a 3D-structure for the frur from the DNA and amino acid sequence and compared it to the structures of other repressors. The programs used in this study were ClustalW, Vectorette II, ORF finder, ExPASy Translate tool, BLASTP, InterProScan, PSIPRED, SWISS-MODEL and DeepView. The results showed rather good homology to other repressors, e.g. purine repressor. Introduction Heterofermentative lactic acid bacteria (LAB) utilize glucose in submerged cultures and produce lactate and ethanol into the culture medium. The ethanol production oxidizes NADH to the form that is needed for the cells to utilize glucose. This oxidization can also be done using other reactions preferably with commercial interest, for instance the reduction of fructose to mannitol. Fructose, however, can also be utilized by the LAB in the glucose pathway by certain enzymes produced when fructose is present in the culture medium, namely fructokinase and glucose-6-phosphate isomerase. The expression of these proteins is regulated by a fructose repressor (FruR). This repressor binds to the 1

2 DNA operator area, when no inducing agent is present in the medium. When an inducer, e.g. fructose is present, the repressor detaches from the DNA and RNA polymerases can begin the synthesis of mrna coding fructokinase and glucose-6-phosphate isomerase. The utilization of fructose in the LAB metabolic activities is not preferable in a process aiming at efficient bioconversion of fructose to mannitol and thus the DNA bound state of the repressor is preferred. As fructose tends to detach the repressor, an anti-inducer with a greater affinity to the enzyme than fructose has, would be valuable. The aim of this study is to compare the structure of the fructose repressor to other repressors and to use databases and computer programs to find out the possible 3D-structure of this enzyme based on its DNA sequence. Methods Cloning and sequencing of fructose operon from Lactobacillus fermentum NRRL-B-1932 Closely related fructokinase sequences from the bacteria Lactococcus lactis sp. lactis, Lc. lactis sp. cremoris, Streptococcus mutans, Pediococcus pentosaceus and Bacillus subtilis were identified using the BLAST search program on the NCBI server. Conserved regions of the fructokinases were then aligned using ClustalW with default settings. Degenerated primers were designed against the most conserved segments of alignment as described by Bartl (1997). A 0.7 kb fragment of L. fermentum fructokinase gene was amplified with PCR using degenerated primer pair, subcloned into cloning vector pgem-t Easy. Plasmids were sequenced by the University of Helsinki Institute of Biotechnology. The Vectorette II system (Sigma Genosys, Ltd., UK) was used to create a genomic library of L. fermentum. Chromosomal DNA of L. fermentum was digested either with BamHI, ClaI, EcoRI or HindIII (New England Biolabs, Beverly, MA), and ligated with 2

3 the respective Vectorette units. Vectorette amplicons were amplified with PCR using sequence specific and Vectorette unit specific primers, subcloned into cloning vector pgm-t Easy, and sequenced. ORF analysis and translation to amino acid sequences Sequence fragments were joined together and analyzed by GCG Wisconsin Package. Open reading frame (ORF) analyses were done with ORF-finder which identifies all possible ORFs in a DNA sequence by locating the standard and alternative stop and start codons. All possible ORFs were translated to amino acid sequences by Expasy Translate tool. Homology and function search The translated amino acid sequence was used in a BLASTP search against Swiss-Protand PDB-databases. The BLAST algorithm is a heuristic search method that seeks words of length W (default = 3 in BLASTP) that score at least T when aligned with the query and scored with a substitution matrix. Words in the database that score T or greater are extended in both directions in an attempt to find a locally optimal ungapped alignment or HSP (high scoring pair) with a score of at least S (alignment score) or an E value (number of hits with a score equal to or better than S that would be expected by chance (the background noise) when searching a database of a particular size) lower than the specified threshold. HSPs that meet these criteria will be reported by BLAST, provided they do not exceed the cutoff value specified for number of descriptions and/or alignments to report. 3

4 Protein function and structure can be determined by comparing so called conserved areas of the amino acid sequence. These are stored in databases e.g. PROSITE, which contains information on over 1000 protein families and domains. There are also programs that do database searches on many databases to find the same information on a broader basis, e.g. InterProScan, which is a European search engine using PROSITE, SwissProt and TrEMBL. In this study we used InterProScan. The program used the amino acid sequence of the FruR to search for similar structures and functions in database proteins. Secondary structure prediction Secondary structure prediction was done by PSIPRED protein structure prediction server. It is a simple and reliable secondary structure prediction method, incorporating two feedforward neural networks which perform an analysis on output obtained from PSI-BLAST (Position Specific Iterated - BLAST). 3D-structure prediction 3D-structure prediction was done by SWISS-MODEL server using the first approach mode. SWISS-MODEL is a server for automated comparative modeling of 3D protein structures. It pioneered the field of automated modeling starting in 1993 and is the most widely-used free web-based automated modeling facility today. Template selection, alignment and model building are done completely automated by the server. The first approach mode provides a simple interface and requires only an amino acid sequence as input data. The server will automatically select suitable templates. Optionally, the user can specify up to five template structures, either from the ExPDB library or uploaded coordinate files. The automated modeling procedure will start if at 4

5 least one modeling template is available that has a sequence identity of more than 25% with the submitted target sequence. However, users need to be aware that the model reliability decreases as the sequence identity decreases and that target-template pairs sharing less than 50% sequence identity may often require manual adjustment of the alignment. In the alignment mode the modeling procedure is initiated by submitting a sequence alignment. The user specifies which sequence in the given alignment is the target sequence and which one corresponds to a structurally known protein chain from the ExPDB template library. The server will build the model based on the given alignment. The project mode allows the user to submit a manually optimized modeling request to the SWISS-MODEL server. The starting point for this mode is a DeepView project file. It contains the superposed template structures, and the alignment between the target and the templates. This mode gives the user control over a wide range of parameters, e.g. template selection or gap placement in the alignment. Furthermore, the project mode can also be used to iteratively improve the output of the first approach mode Modeling procedure (Adapted from Schwede et al. (2003)) All homology-modeling methods consist of the following four steps: (i) template selection; (ii) target template alignment; (iii) model building; and (iv) evaluation. These steps can be iteratively repeated, until a satisfying model structure is achieved. Several different techniques for model building have been developed. The SWISS-MODEL server approach can be described as rigid fragment assembly, which will be outlined briefly. The SWISS-MODEL server template library ExPDB is extracted from the PDB. In order to allow a stable and automated workflow of the server, the PDB coordinate files are split into individual protein chains and unreliable entries, e.g. theoretical models and low quality structures providing only C coordinates, are removed. Additional information 5

6 useful for template selection is gathered and added to the file header, e.g. probable quaternary structure, quality indicators like empirical force field energy or ANOLEA mean force potential scores. To select templates for a given protein, the sequences of the template structure library are searched. If these templates cover distinct regions of the target sequence, the modeling process will be split into separate independent batches. Up to five template structures per batch are superposed using an iterative least squares algorithm. A structural alignment is generated after removing incompatible templates, i.e. omitting structures with high C root mean square deviations to the first template. A local pair-wise alignment of the target sequence to the main template structures is calculated, followed by a heuristic step to improve the alignment for modeling purposes. The placement of insertions and deletions is optimized considering the template structure context. In particular, isolated residues in the alignment ( islands ) are moved to the flanks to facilitate the loop building process. To generate the core of the model, the backbone atom positions of the template structure are averaged. The templates are thereby weighted by their sequence similarity to the target sequence, while significantly deviating atom positions are excluded. The template coordinates cannot be used to model regions of insertions or deletions in the targettemplate alignment. To generate those parts, an ensemble of fragments compatible with the neighboring stems is constructed using constraint space programming (CSP). The best loop is selected using a scoring scheme, which accounts for force field energy, steric hindrance and favorable interactions like hydrogen bond formation. If no suitable loop can be identified, the flanking residues are included to the rebuilt fragment to allow for more flexibility. In cases where CSP does not give a satisfying solution and for loops above 10 residues, a loop library derived from experimental structures is searched to find compatible loop fragments. The reconstruction of the model side chains is based on the weighted positions of corresponding residues in the template structures. Starting with conserved residues, the model side chains are built by iso-sterically replacing template structure side chains. 6

7 Possible side chain conformations are selected from a backbone dependent rotamer library, which has been constructed carefully taking into account the quality of the source structures. A scoring function assessing favorable interactions (hydrogen bonds, disulfide bridges) and unfavorably close contacts is applied to select the most likely conformation. Deviations in the protein structure geometry, which have been introduced by the modeling algorithm when joining rigid fragments are regularized in the last modeling step by steepest descent energy minimization using the GROMOS96 force field. Empirical force fields are useful to detect parts of the model with conformationalerrors. In our own experience and the work of others, energy minimization or molecular dynamics methods are in general not able to improve the accuracy of the models, and are used in SWISS- MODEL only to regularize the structure. However, the successful application of restricted molecular dynamics for improving homology models has recently been reported for a few test cases. To derive more general rules of engagement of molecular dynamics, further systematic experiments have to be conducted. The four modeling steps template superposition, target-template alignment, model building and energy minimization have been implemented in the program ProModII in ANSI C. Comparative study Protein Data Base (PDB) The Protein Data Base (PDB) is a single worldwide repository for the processing and distribution of 3D-structure data of large molecules of proteins and nucleic acids. Repressor protein 3D-structures were downloaded from the PDB database to obtain model structures that can be compared with the frur model calculated by SWISS- MODEL. 7

8 DeepView The protein model was visualized and compared to other repressor structures with the DeepView program. The program DeepView (Swiss-PdbViewer) was designed to integrate functions for protein structure visualization, analysis and manipulation into a sequence-to-structure workbench with a user-friendlyinterface. It allows the user to manage complex modeling projects and is publicly available from the ExPASy server. With DeepView one can search for suitable modeling templates and download the corresponding PDB or ExPDB files directly from the DeepView server. Using the integrated sequence alignment tools and structural superposition algorithms, a target sequence can be mapped onto the modeling templates in one step. Then the initial sequence alignment can be optimized manually while the anticipated changes in the model backbone are reflected in real-time in the displayed structural superposition. The complete project file is then submitted to the SWISS-MODEL server for model building. The resulting protein model can be visualized and analyzed using the integrated tools in DeepView program. Results Analysis of the amino acid sequence The ORF finder found three open reading frames. The amino acid sequence analysis by BLAST showed that the operon consisted of two different genes coding fructokinase and phoshoglucose isomerase and their regulator protein frur. The analysis resulted in a too short frur sequence; the HTH motif was left away. This might have occurred because of the repressors transcription initiation codon TTG which was not recognized by the ORF finder program. In the literature we found that the HTH-motif was essential for a 8

9 functional repressor protein. It is the region that binds to DNA. The DNA sequence was then extended so that the HTH-motif coding region was also included in the frur sequence before further analysis. The ExPASy Translate tool found only one ORF that could code the amino acid sequence of the frur. This sequence was used for the rest of the study. Homology and function The InterProScan search result is shown in Figure 1. The program identified the beginning of the amino acid sequence as a Helix-Turn-Helix (HTH) motif, which classifies the protein as a LacI family protein. The LacI family consists of various regulatory proteins, such as ascg, ccpa, cytr, galr, laci, opnr, purf and scrr (Adhya and Weickert, 1992). The HTH motif is situated towards the N-terminus in the proteins of this family. Secondary structure prediction The predicted secondary structure obtained from PSIPRED is shown in Figure 2. The green tubes represent α-helixes, yellow arrows β-sheets and strands loop areas. The columns above the structure represent the probability of the structure. 3D-structure prediction The predicted 3D-structure of frur is shown in Figure 3. The α-helixes are shown in red and β-sheets in yellow. The blue-gray areas are loops or uncertainties of the model. The 9

10 frur has a headpiece containing three α-helixes (helix-turn-helix motif and hinge helix motif). The most probable active site is surrounded by the N-subdomain and C- subdomain, both containing α-helixes and β-sheets as well as uncertain structure. Figure 1. InterProScan results from the analysis of the frur amino acid sequence. Comparative study Figure 4 shows the surface models of frur and purr. The α-helixes are shown in red and β-sheets in yellow. Loops and uncertain areas of the model are shown in gray. Structural similarity of the two repressor proteins can clearly be observed. 10

11 Figure 2. The predicted secondary structure of the frur. The green tubes represent α-helixes and yellow arrows β-sheets. 11

12 Figure 3. SWISS-MODEL graphical representation of the frur. The red areas represent α-helixes and yellow areas β-sheets. The blue-gray areas are loops or uncertainties of the model. Figure 4. Surface models of the fructose repressor (left) and purine repressor (right). The purine repressor has a DNA strand attached to the headpiece. The red areas represent α-helixes and yellow areas β-sheets. Gray areas are loops or uncertainties of the model. 12

13 Discussion The predicted 3D-structure of the L. fermentum frur repressor protein still has many uncertain areas that will need more accurate modeling. Loop structures are normally very difficult to model, as they are quite flexible structures. Certain amino acids, like glycine and alanine, that are very small, can move easily and thus can not form rigid structures compared to the bigger amino acids, e.g. proline. The next step in modeling could be comparing the automatically predicted repressor model to known 3D-structures manually, one amino acid at a time, by using a superposition program. This can be done in a protein modeling program, e.g. Quanta. The model obtained in this work shows that this frur is a repressor with a typical DNA binding site. The model of the active site is still too uncertain for even guessing suitable inducers or anti-inducers that could be used for improvement of mannitol yield from fructose. References Adhya, S. and Weickert, M.J., A family of bacterial regulators homologous to gal and lac repressors, J. Biol. Chem. 267 (1992) BLAST, Bartl, S., Amplification using degenerate primers with multiple inosines to isolate genes with minimal sequence similarity. In: Methods in molecular biology, Vol 67: PCR cloning protocols: from molecular cloning to genetic engineering, Ed. B.A.White, Humana Press Inc., Totowa, NJ 1997, p

14 ExPASy Translate tool, InterProScan, ORF-finder, PROSITE, Protein Data Bank, PsiPred, Schwede, T., Kopp, J., Guex, N., and Peitsch, M.C., SWISS-MODEL: an automated protein homology-modeling server, Nucl. Acids Res. 31 (2003) SwissModel,