千里光S一腺苷甲硫氨酸合成酶（SAMS）的結構域與功能位點分析

譚昊 文春菊 錢倩 錢剛

摘要：基于前期研究工作中構建的千里光全長cDNA文庫，我們分離到千里光s-腺苷甲硫氨酸合成酶（s-adenosylmethionine synthase，.SA MS）基因，本研究分析了該基因開放閱讀框，并對該多肽3個高度保守序列形成的結構域和功能位點之間的關系進行了探討。結果表明，該基因（GenBank ID：KC149908.1）編碼的蛋白質由394個氨基酸殘基組成，理論等電點5.48，相對分子質量43.40 kD；結構域比對和3-D模型分析結果表明，a-helix/盧-sLrand是該蛋白結構域的主要組分，且SPOUT結構與甲基轉移酶反應密切相關，其功能可能涉及核酸、蛋白質和磷脂的合成，本研究揭示SAMS的氨基酸保守基序是形成高級結構并決定其生物學功能之關鍵所在。

關鍵詞：結構域；三維模型比對；S-腺苷甲硫氨酸合成酶（SAMS）；千里光

中圖分類號：Q949.783.5 文獻標識碼：A 文章編號：1007-7847（2015）03-0203-07

1 Introduction

Regulation of the tetrapyrrole biosynthesispathway is complex and involves several regulatorysystems. Protein function can be thought of on dif-ferent interdependent levels and may be divided in-to three major categories： molecular function， bio～logical process and cellular component[1]. S-adenos-ylmethionine synthase （SAMS）， the second mostprevalent enzyme substrate in cells after ATP， isthe major methyl donor for essential methylation re-actions and serves as a substrate in polyaminebiosynthesisc[2]. In addition to its role in radical SAMSenzymes， it transfers one electron from an iron-sul-fur cluster to the SAMS cofactor， which is thencleaved into methionine and a highly oxidizing radi-cal[3]. As the biological function of protein moleculeis accurately described by its three-dimensional st-ructure， protein-fold structural domains and inter-acting components in whole metabolic networkst[4，5]，ones of the most common motivations for predictingthe protein structure， are used to gain insight intothe protein's biological function. It is neverthelessan efficient process in the biosynthesis reaction andRNA transcription termination in， vitro[6]； despite theheavy metabolic demands， which vary according tothe changeable gene sites and affect functional rolesof this pathway， no direct evidence is available toclarify SAMS functional assay from relationship be-tween the key residues of highly conserved motifsand the protein-fold structural domams.

Senecio scandens Buch.-Ham. ex D. Don， pre-dominantly selling annual， plays an important rolein anti-microorganism involved in Chinese tradi-tional medicinal plant and has a widespread distri-bution in a few ecological habitats of China[7]. Owingto its important antibacterial source in Chinese tra-ditional medicine， the biological features should bedistinguished at the molecular level to facilitatebreeding， gene discovery or industrial applications.As a general trend， the biological usefulness of thepredicted protein models relies on the accuracy ofthe structure prediction[8]， although a structural ins-ight into the arrangement of the components in suchcomplexes is still limited[9]. Recent advances in co-mputer algorithms for predicting protein structureand function have alleviated this problem and pro-vide biologists with valuable information about theirproteins of interest[10] Therefore， we here focus on： 1）clarifying SAMS functional sites determined by itshighly conserved motifs； 2） presenting 3-D modelalignments for a better understanding of the struc-tural correlations to functional roles， 3） elucidatingthe phylogenetic relationship of the S-Adenosylme-thionine in the high plants.

2 Materials and methods

2.1 Plant materials

The experimental materials Senecio scanden，swere harvested from the diverse eco-geographic re-gions of Yunnan-Guizhou plateau. In this study， theelite antibacterial sample （SC-36） with the highquality of antibacterial feature was selected to con-struct cDNA library according to the methods ofShapiro and Baneyx f1u using a series of standard-ization bacteria involving Staphylococcus aureu.s，Pseudomonas ae -ru.ginosa， Escherichia coli，Salmonella paratyphi， Shigella flexneri， Aeromonassobria， and Edwardsiella tarda

2.2 Construction of full-length cDNA libraryand sequence data trimming

Leaf tissue of the experimental seedlings washarvested for RNA extraction， using TRIzol-RNATotal RNA Isolation Kit （Invitrogen， China）.SMART cDNA library construction kit was appliedto generate a full-length cDNA library according tothe manufacture's suggestions. The ligation product（5 μL） of the resultant double cDNA and the vectorpDNR-UB was transferred to electrocompetent cellXLl-Blue （25 μL）. The plasmid DNA of each clonewas directly prepared from bacterial cultures of aglycerol stock plate by the RCA method using aTempliPhi HT DNA amplification kit （GE Health-care， UK）. End sequencing of 10 000 clones wascarried out with iCycler iQ SYBR Green PCR（BIO～RAD Co.， LTD.， USA） using M13 sense andantisense primer. Raw sequence data （chro-matograms） were base-called using the Phred pro-gram and vector sequences were then detected byusing cross-match. The low quality region （Phredquality score < 20， and more than > 20 bases re-peated） was cliscarded. We trirumed off the vectorsequences of both encls of each read using the sim4program[7] Sequences data of'lengths shorter than 100bases after the trimming process were also omittedfor further analysis.

2.3 Sequence alignment and phylogenetic an-alysis

The cDNA sequence prediction was conductedwith GenScan software （http：//genes.mit.edu/GEN-SCAN.html）. Sequence similarity analysis in Gen-Bank was performed using the Blast 2.1 search tool（http：//www.ncbi.nlm.nih.gov/blast/）. ClustalW soft-ware （http：//www.ebi.ac.uk/clustalw/） was used for al-ignment of multiple sequences. Identified ORFs ofone transcript （SAMS） was translated into aminoacid sequences， and multiple alignments of deducedamino acid sequences were performed usingClustalW with default optionsnz. Nucleotides and aminoacid sequence analyses were performed with DNA-MAN program. Phylogenetic trees and molecularevoluLionary analyses were constructed based on thebootstrap Neighbor-joining （NJ） method with aJukes-Cantor model for DNA sequences and Pois-son correction model for amino-acid sequences byMEGA v4.O[13]. The stability of internal nodes was as-sessed by bootstrap analysis with 1 000 replicates.

2.4 Prediction and functional assays of proteinmolecule

SAMS was selected to perform further bioinfor-matics analysis according to the methods of Umeza-wa et al [14]. Signature amino acid patterns for SAMsynthetases were retrieved from the PROSITEdatabase of protein families and domains. The se-quences excluded in all searches were submitted tothe InterPro version 4.2 with DBrelease12.1 to iden-tify their functional domains. To predict the bio-physics characteristics of the putative protein ofSAMS， software on the ExPASy Proteomics Server（http：//au.expasy.org/） was used. SignalP-4.0 soft-ware was applied to analyze the protein signal pep-tide （http：//www.cbs.dtu.dk/services/SignalP/）. Theprediction and analysis for the protein structuraldomain and functional site were finished usingPROSITE software （http：//www.expasy.org/prosite/）.The 3-D shape of the putative protein conservativedomain was performed with the 3 ～D ConservativeDomain Architecture Retrieval Tool of Blast （http：//swissmodel.expasy.org/）， and its alignment model wasobtained from Database of VAST model （http：//www.ncbi.nlm.nih.gov/blastj）.

3 Results

3.1 Sequence characteristics and molecularevolutionary of SAMS

Here， the isolation of a cDNA encoding SAMSis obtained from a full-length cDNA library inSenecio scandens Buch.-Ham. ex D. Don. The pre-sent gene （GenBank ID： KC149908.1） encodes aprotein composed of 394 amino acid residues， withthe theoretical isoelectric point of 5.48 and thepredicted molecular weight of 43.40 kD. As shownin Fig.l. 47 different genera are accepted to presentthe phylogenetic tree depending on these selectionsof the highest scores of E-values in the samespecies. Based on the deduced amino acid sequenceof SAM， a combined phylogenetic tree reveals thatthe present accession （Senecio scan，dens Buch.-Ham. ex D. Don） has the closest genetic relation toPopulus trichocarpa among the selected species.

3.2 Deternunation on conserved sequences

As a result of sequence alignments by runninga BlastN search against the GenBank "nr/nt"databases， a complete coding sequence of SAMgene is selected to perform with sequencing analy-sis. Seven representative accessions are further ap-plied to determine the conservative motif sequencesdepending on Lhe most diverse genetic distance. Asshown in Fig.2， the amino acid sequence compari-sons validate the correclness of the current classifi-cation of the SAMS domain protein， sharing 92.710/ohomology with these of the high plant counterparts.Accordingly， three conservative motifs are furtherused to observe the relationship between structuraldomains and functional sites， included in N-termi-nal conserved region， M-conserved region and C-terminal conserved region.

3.3 StructuraIRy similar alignments of conser-vative domain in SAMS

Next， the 3-D conformation of' the putat.ive do-main of SAMS protein is applied to determine thefunccional site linkage to structural domain of thoseconservalive residues， using the DALI-server. Justas demonstrated by the alignments of' 3-D shape7the highly consistent structural distributions arefouncl between SAMS of Ricinus communiS（XP_002512570.1） and the present protein， involvedof' N-terminus （Fig.3A）， M-conserved region （Fig.3B）and C-terminus （Fig.3C）.

3.4 The topology prediction on SAMS

The Lopology prediction also observes that theminimal conservative core contains the so -calleclSPOUT-domain （related to the SpoU- and TrmD-melhyllransferases） in Senecio scan，dens Buch. -is the other deCining feature of the SPOUT-domainfold. Moreover. the knol region is stabilized by anexlended network of hydrophobic interactions thatform a genuine hydrophobic core.

4 Discussion

When a gene has been identified， there mayHam. ex D. Don （SC-36）. The result shows that theSAMS topology structure is comprised of six com-pactly folded domains of“α-helix/β-strand sheetsand a long protruding globular loop （Fig.4A）. In thiscase， a relaxed α-helix/β-strand subunit is insertedinto the core fold of the SPOUT-domain （Fig.4B）.As a result of this observation， a deep trefoil knot isinvolved in both the formation of the co-factor-binding sile and part of the dimer interface， whichbe more than one polymorphism within these ortho-logues from different species. A muLation causing aradical change in Lhe amino acid is more likely toaffect the properties of the protein than a conserva-tive amino acid substitution[15]. Some of the most suc-cessful approaches use a phylogenetic Lree to rankthe residues by evolutionary importance and thenmap this ranking onto a structure if one is availablec16l.Here， our results indicate that a stable conformationfrom the key residues of highly conserved motifs islikely to be needed for protein function acrossspecies. Bioinformatics analysis shows that these re-gions in three conserved units of SAMS protein arehighly identical sequences among the differentspecies （Fig. 1 and Fig.2）， based upon the target-template sequence alignment of those accessions.Thus， the greatest degree of consensus residues hasconfidently been used to clarify the structural corre-lations to some potential functional sites， because oftheir similar sequences and secondary structures.

As proteins from different evolutionary originsmay have similar structure， threading methods aredesigned to match the query sequence directly ontothe 3-D structures of other solved proteins， with thegoal of recognizing folds similar to the query evenwhen there is no evolutionary relationship betweenthe query and the template proteinuo. Structural datacan be used to detect proteins with similar functionwhose sequences have diverged beyond a level sim-ilarity that can reliably detected using sequencecomparison methods[1]. Therefore， our target-to-te-mplate model alignment （Fig.3） will also provide anaccurate structural comparison on 3-D model tem-plate from the most contrasting diverse species，even if there is no evolutionary relationship betweenour present target sequenc.e and the template pro-tein. After the major model alignments have beenmade， the level of conservative sequence oloservedfor the residues strongly suggests a functional im-portance for these armno acids in structural do-mains. As expected for template-based homologymodels， structure-based domain alignments of SAMSproteins indicate that the crucial functional sites aredominated by three conserved motifs and none ofthe mutated amino acid positions is in the vicinityof the SAMS-binding sites.

Interestingly， 3 -D alignments of protein-folddomains in this scudy show that the highly struc-tural similarity results likely in the same and/orsimilar function in proteins. Just as shown in Fig.4，a genuine hydrophobic core from the core fold ofthe SPOUT-domain may be the key methyl groupdonor for the methyltransferase reactions involvingDNA， RNA， proteins， and phospholipids. As shownin Fig.3 and Fig.4， the other positively charged areainvolves residues of the irregularly structured sur-face loop which is close in space to the cluster ofconservative Arg （from R278 to I303）. The irregularlystructured surface loop is poorly conservaLive， andits utilization would be in agreement with the obser-vation that in many other SPOUT-methyltransferas-es insertion or extension elements are exploited asauxiliary RNA-binding elements[18]. Taylor et al.[19]. asparallel with our results， also found that the preva-lence of backbone-mediated interactions with theligand correlates with the lack of sequence conser-vation for armno acids in the SAMS-binding pock-ets between Nepl and other members of the SPOUT-class of methyltransferases. The r.opology and thelocation of the co-factor-binding site are exactlyconservative in SPOUT-domain even if the cofactoradopts an extended conformation in the classicalRossmarit-fold methyltransferases[20].

In coriclusion， phylogenetic tree and sequencealignment are applied to detect highly conservedmotifs of SAMS protein for observing the relalion-ship between structural domains and func，tionalsites. Both 3-D model and structural domain align-ments indicate that a genuine hydrophobic corefrom SPOUT-class may perform with the methyl-transferase reactions， which composed of the keyresidues of highly conserved motifs of SAMS pro-tein. This work sheds light on the functional sitesare attributed to the structural domains in S-adeno-sylmethionine synthase.