Unique protein architecture of alanyl-tRNA synthetasefor aminoacylation, editing, and dimerizationMasahiro Naganumaa, Shun-ichi Sekinea,b, Ryuya Fukunagaa,1, and Shigeyuki Yokoyamaa,b,2

aDepartment of Biophysics and Biochemistry, Graduate School of Science, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan; and bRIKENSystems and Structural Biology Center, Yokohama Institute, 1-7-22 Suehiro-cho, Tsurumi-ku, Yokohama 230-0045, Japan

Edited by Paul R. Schimmel, The Scripps Research Institute, La Jolla, CA, and approved April 6, 2009 (received for review February 17, 2009)

Alanyl-tRNA synthetase (AlaRS) specifically recognizes the majoridentity determinant, the G3:U70 base pair, in the acceptor stem oftRNAAla by both the tRNA-recognition and editing domains. In thisstudy, we solved the crystal structures of 2 halves of Archaeoglo-bus fulgidus AlaRS: AlaRS-�C, comprising the aminoacylation,tRNA-recognition, and editing domains, and AlaRS-C, comprisingthe dimerization domain. The aminoacylation/tRNA-recognitiondomains contain an insertion incompatible with the class-specifictRNA-binding mode. The editing domain is fixed tightly via hydro-phobic interactions to the aminoacylation/tRNA-recognition do-mains, on the side opposite from that in threonyl-tRNA synthetase.A groove formed between the aminoacylation/tRNA-recognitiondomains and the editing domain appears to be an alternativetRNA-binding site, which might be used for the aminoacylationand/or editing reactions. Actually, the amino acid residues requiredfor the G3:U70 recognition are mapped in this groove. The dimer-ization domain consists of helical and globular subdomains. Thehelical subdomain mediates dimerization by forming a helix–loop–helix zipper. The globular subdomain, which is important forthe aminoacylation and editing activities, has a positively-chargedface suitable for tRNA binding.

crystal structure � dimerization domain � aminoacyl-tRNA synthetase �proofreading � wobble base pair

Aminoacyl-tRNA synthetases (aaRSs) catalyze the ligation ofcognate amino acids and tRNAs, and thus establish thegenetic code in protein biosynthesis. They are modular proteinscomposed of an aminoacylation domain and a few additionaldomains for discrete functions, such as tRNA binding, oligomer-ization, and amino acid proofreading (1, 2). The 20 aaRSs aredivided into 2 classes, I and II, based on the 2 unrelated types ofaminoacylation domains (3, 4). The aminoacylation reactionoccurs at the catalytic site on the aminoacylation domain, andthe reaction generally consists of 2 steps: the initial activation ofthe amino acid with ATP to generate the aminoacyl-adenylate,followed by the transfer of the aminoacyl moiety to the 3� endof the tRNA. Although the aminoacylation is generally accurate,several aaRSs cannot completely avoid the misactivation of anoncognate amino acid, when it is similar to the cognate one. Tosolve this problem, these aaRSs use a proofreading mechanism,in which the incorrect products are hydrolyzed at the active sitein the editing domain.

Alanyl-tRNA synthetase (AlaRS) is one of the class II aaRSsand consists of 4 domains: the N-terminal class II aminoacylationdomain, the tRNA-recognition domain, the editing domain, andthe C-terminal oligomerization (dimerization or tetrameriza-tion) domain (Fig. 1A) (1, 2). AlaRS occupies a special positionin the history of aaRS research. Escherichia coli AlaRS wasamong the first aaRSs that were cloned, sequenced, and char-acterized genetically and biochemically (1, 5, 6). tRNAAla con-serves a unique G3:U70 wobble base pair in the acceptor stem,and this base pair dictates the tRNA identity toward AlaRS(7, 8). This remarkable finding, that a small number of nucleo-tide residues serve as the predominant determinant for thetRNA identity, accelerated the search for the identity determi-

nants of other aaRS–tRNA pairs. It was also striking that thepredominant identity determinant of a tRNA exists in theacceptor–stem duplex, rather than the anticodon and the dis-criminator base (9, 10). Actually, AlaRS can aminoacylate small,isolated portions of tRNA, such as a ‘‘minihelix’’ and a ‘‘micro-helix,’’ as long as they have the G3:U70 base pair (11). TheG3:U70 base pair is considered to be recognized from the minorgroove side (12, 13).

An E. coli AlaRS fragment comprising the aminoacylation andtRNA-recognition domains (the N-terminal 461 residues) canspecifically aminoacylate tRNAAla (1). The crystal structure ofthe corresponding fragment (AlaRS-N) from the bacteriumAquifex aeolicus was reported (14, 15). It revealed that AlaRSdoes not dimerize through the aminoacylation domain, in con-trast to the other class II aaRSs. The structures of amino acid-and ATP-bound AlaRS-N revealed how the cognate alanine andthe noncognate glycine and serine interact with the aminoacy-lation site. AlaRS is one of the aaRSs that use the proofreadingmechanism, in that mischarged products, such as Gly-tRNAAlaand Ser-tRNAAla, are transferred to the editing domain, wherethe ester bond is hydrolyzed (2). A defect in the AlaRS editingactivity causes cell death in the mouse nervous system (16). It wasrecently reported that the E. coli AlaRS editing domain pos-sesses a region, distinct from the N-terminal domains, thatrecognizes the G3:U70 base pair (17). Therefore, AlaRS maytransfer the acceptor stem of tRNAAla from the first binding sitein the aminoacylation domain to the second site in the editingdomain, in contrast to the other editing aaRSs (classes I and II),which have been proposed to shuttle the flexible single-strandedCCA terminus of the tRNA between the aminoacylation andediting catalytic sites (18–22). The C-terminal domain of AlaRSis not only essential for the oligomerization, but also importantfor the aminoacylation and editing reactions (17, 23). Smallproteins homologous to the AlaRS editing domain, designatedAlaX, are found in many organisms (24, 25). They are active inthe trans hydrolysis of misacylated tRNAAla in vitro (24). Thecrystal structures of AlaX-S (specific to Ser-tRNAAla) andAlaX-M (specific to Ser-tRNAAla and Gly-tRNAAla) from thearchaeon Pyrococcus horikoshii have been reported (26, 27).

The structures of the editing and oligomerization domains, thebasis of oligomerization, and the domain arrangement in thefull-length AlaRS have remained elusive. We previously suc-ceeded in the crystallization of 2 fragments of AlaRS from the

Author contributions: S.Y. designed research; M.N., S.S., and R.F. performed research; M.N.,S.S., and S.Y. analyzed data; and M.N., S.S., and S.Y. wrote the paper;.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Data deposition: The coordinates and structure factors have been deposited in the ProteinData Bank, www.pdb.org (PDB ID codes 2ZTG and 2ZVF).

1Present address: Department of Biochemistry and Molecular Pharmacology, University ofMassachusetts Medical School, Worcester, MA 01605.

2To whom correspondence should be addressed. E-mail: yokoyama@biochem.s.u-tokyo.ac.jp.

This article contains supporting information online at www.pnas.org/cgi/content/full/0901572106/DCSupplemental.

www.pnas.org�cgi�doi�10.1073�pnas.0901572106 PNAS � May 26, 2009 � vol. 106 � no. 21 � 8489–8494

BIO

CHEM

ISTR

Y

Dow

nloa

ded

by g

uest

on

June

15,

202

1

http://www.pnas.org/cgi/content/full/0901572106/DCSupplementalhttp://www.pnas.org/cgi/content/full/0901572106/DCSupplemental

archaeon Archaeoglobus fulgidus, AlaRS-�C, comprising theaminoacylation, tRNA-recognition, and editing domains, andAlaRS-C, comprising the dimerization domain (23). In thepresent study, we determined their crystal structures at 2.2- and3.2-Å resolutions, respectively. The AlaRS-�C structure re-vealed a unique arrangement of the editing domain, relative tothe aminoacylation/tRNA-recognition domains, and the ar-chaea-specific insertions/deletions. The AlaRS-C structure pro-vided the basis of dimerization, via the formation of a helix–loop–helix zipper (HLHZ). The structures suggested the domainorganization in the full-length AlaRS dimer, and thus couldserve as a platform for future analyses of how the aminoacyla-tion/tRNA-recognition domains and the editing domain of

AlaRS independently recognize the G3:U70 base pair oftRNAAla.

ResultsStructure Determination. A. fulgidus AlaRS is a homodimer of 906amino acid residue polypeptides (23). It was genetically dividedinto 2 parts, AlaRS-�C (residues 1–739) and AlaRS-C (residues736–906) (23), and both structures were solved (Table S1).AlaRS-�C is composed of the class-II specific aminoacylationdomain, the tRNA-recognition domain, and the editing domain.The structure of AlaRS-�C complexed with an alanyl-adenylateanalog, 5�-O-[N-(L-alanyl)sulfamoyl] adenosine (Ala-SA), wasdetermined at 2.2 Å (Fig. 1B). The crystallographic asymmetricunit contained 1 AlaRS-�C molecule. The refined model has Rand Rfree factors of 21.5% and 26.4%, respectively. AlaRS-Ccomprises the dimerization domain, and its structure was deter-mined at 3.2 Å (Fig. 1C). Two AlaRS-C molecules form ahomodimer, and there are 4 dimers in the asymmetric unit. Therefinement converged to R and Rfree factors of 20.5% and 27.6%,respectively (Table S1).

The Aminoacylation Domain. The A. fulgidus AlaRS-�C structurerevealed the aminoacylation domain (residues 1–257), composedof a central antiparallel �-sheet (�1–�8 and �10) and 5 �-helices(�1–�5), which is typical of the class-II aaRSs. It is superposableon that of A. aeolicus AlaRS-N, with an rmsd of 1.9 Å for 179C� atoms. A. fulgidus AlaRS possesses an archaea-specificN-terminal extension (AddA1, residues 1–58), including �1, �2,�1, and �2 (Fig. 2 and Fig. S1 A). �1 and �2 are integrated in thecentral �-sheet, and thus AddA1 is part of the aminoacylationdomain. Although the bacterial and eukaryal AlaRSs lackAddA1, they instead possess 2 insertions (depicted as InsB/E1and InsB/E2 in A. aeolicus AlaRS), which occupy the corre-sponding space. A. fulgidus AlaRS also contains an insertion(InsA1, residues 226–232) including �9 (Fig. 2 and Fig. S1 A).Lys-229, at the tip of InsA1, seems to occupy the position ofLys-73 in the E. coli enzyme, which cross-links to tRNAAla (28).

A clear electron density corresponding to Ala-SA is visible inthe active-site cleft (Fig. 1B and Fig. S2). The manner ofinteraction with Ala-SA in the aminoacylation active site isdescribed in SI Text.

The tRNA-Recognition Domain. The �-helix-rich middle domain ofAlaRS-�C (residues 258–484), which is supposed to interactwith the tRNA acceptor arm, is composed of 11 �-helices and a2-stranded short parallel � sheet (Fig. 1B and Fig. S1 A). Thisdomain can be divided into 2 subdomains, designated here as

A

B Aminoacylation

tRNA recognitionMid2

Mid1

N

C

C

N

Editing

Linker

Globular subdomains

Helical subdomains

N

C

Editing core

β barrel

InsA2

AddA1

Helix-loop

InsA1

motif1

C

Aminoacylation tRNA recognition Editing Oligomerization

AlaRS

AlaRS-ΔC

AlaRS-C

1 906

1 739

736 906

Fig. 1. The structures of AlaRS-�C and AlaRS-C. (A) Domain organizations ofA. fulgidus AlaRS, AlaRS-�C, and AlaRS-C. Shown are the aminoacylationdomain (green), Mid1 (blue) and Mid2 (cyan) in the tRNA-recognition domain,the �-barrel (yellow) and editing-core (orange) subdomains in the editingdomain, and the helical (midnight blue) and globular (light blue) subdomainsin the dimerization domain. (B) A ribbon representation of AlaRS-�C. Themodel is colored as in A, and the N-terminal addition (AddA1) and insertions(InsA1 and InsA2) are highlighted in purple and brown, respectively. Motif 1in the aminoacylation domain (violet), the helix–loop in the editing domain(salmon), and the linker connecting the tRNA-recognition and editing do-mains (red) are shown. Ala-SA in the aminoacylation site and the editing-sitezinc ion are depicted as cpk models. (C) The AlaRS-C dimer is shown as a ribbonmodel. One molecule of the dimer is colored as in A, and the other moleculeis colored gray.

InsA1

InsB/E1

InsB/E2

InsA2

AddA1

A.fulgidus A.aeolicusMid2

Mid2

A B

Fig. 2. The aminoacylation and tRNA-recognition domains. (A) The amino-acylation and tRNA-recognition domains of A. fulgidus AlaRS, colored as inFig. 1B, are shown. (B) The A. aeolicus AlaRS-N structure, shown in the sameorientation. The 2 regions missing in A. fulgidus (InsB/E1 and InsB/E2) arecolored brown, and Mid2 is shown in gold.

8490 � www.pnas.org�cgi�doi�10.1073�pnas.0901572106 Naganuma et al.

Dow

nloa

ded

by g

uest

on

June

15,

202

1

http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=ST1http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=ST1http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=SF1http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=SF1http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=SF2http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=STXThttp://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=SF1

Mid1 (residues 258–419) and Mid2 (residues 420–484). Thesesubdomain structures in A. fulgidus AlaRS are similar to theircounterparts in A. aeolicus AlaRS (14), as revealed by the rmsdsof 2.4 and 2.3 Å, respectively. Mid1 tightly contacts the amino-acylation domain by hydrophobic interactions, whereas Mid2protrudes from the rest of the protein body. The �-helix (�13)connecting Mid1 and Mid2 is continuous, whereas the corre-sponding helix is distorted in the middle in A. aeolicus AlaRS-N.Thus, Mid2 in A. fulgidus AlaRS-�C is oriented outward by�20°, compared with that of A. aeolicus AlaRS-N (Fig. S3). It isfurther tilted by �50°, because the last �-helix of Mid2 isconnected to the editing domain by the linker. In the beginningof the subdomain, the Mid1, archaeal AlaRSs possess an inser-tion of �50 amino acid residues, which is missing in the bacterialAlaRSs. In the A. fulgidus AlaRS-�C structure, this insertion(InsA2, residues 277–330) assumes a helix–loop–helix structure(�8–�9) and is bent back to form part of the active-site cleft. Thepresence of this insertion seems to be incompatible with thetRNA interaction manner proposed previously for the bacterialAlaRS (14), as discussed below.

The Editing Domain. The editing domain of A. fulgidus AlaRS(residues 501–737) consists of 2 subdomains, the N-terminal�-barrel subdomain (residues 501–588) and the C-terminal �/�subdomain (the editing core, residues 589–737), composed of 2central �-helices (�17 and �18, residues 590–614 and 642–657,respectively) sandwiched by 3- and 4-stranded antiparallel�-sheets (Fig. 1B and Fig. S1B). The editing domain is super-posable on P. horikoshii AlaX-M [Protein Data Bank (PDB) IDcode 2E1B], with an rmsd of 1.5 Å for 202 C� atoms. The �/�subdomain also superposes well on P. horikoshii AlaX-S (PDBID code 1WXO) lacking the �-barrel subdomain, with an rmsdof 1.4 Å for 134 C� atoms.

A cavity is formed at the subdomain interface. At thebottom, His-600, His-604, His-707, and Cys-703 coordinate ametal ion, which is supposed to be the editing active center(Fig. S4A). The tetrahedral coordination and the intenseanomalous peak observed at the metal site suggested that themetal is a zinc, as in the case of the AlaX proteins. Thr-603,Gln-620, Gln-682, and Gln-701 constitute the cavity wall.His-600, Thr-603, His-604, Gln-620, His-707, and Cys-703 areconserved among the AlaRSs. Glu/Gln occupies the positioncorresponding to Gln-701. Gln-620, Gln-682, and Gln-701correspond to Thr-30, Asp-92, and Asp-114, respectively, in P.horikoshii AlaX-S, which are involved in interactions withserine (Fig. S4B) (27). In AlaX-S, Thr-33 is also involved in theserine interaction, but AlaRS lacks the corresponding residue.The conserved Thr-603 in AlaRS could structurally compen-sate for the absence of this residue. The glycine-rich loop of the�-barrel subdomain resides at the entrance of the cavity andmight interact with 3� end of the tRNA.

Position of the Editing Domain. The editing domain of A. fulgidusAlaRS is connected to the last �-helix of the tRNA-recognitiondomain by a 38-Å-long loop, consisting of 16 amino acidresidues (residues 485–500). The editing domain contacts theaminoacylation domain to form a hydrophobic core (Fig. 3B).Ile-670, Tyr-681, Phe-678, and Val-685, from a helix–loopstructure (residues 669–689) in the editing domain, interactwith Tyr-84 in motif 1 (residues 61–88), Trp-90, Phe-107, andVal-112 of the aminoacylation domain. Arg-89, Trp-619, andArg-679 are stacked. The editing domain also contacts Mid1of the tRNA-recognition domain (Fig. 3C). His-617, Thr-635,and Phe-637 in the editing domain and Val-356, Val-403,Thr-407, Ile-411, and Leu-414 in Mid1 form a hydrophobiccore. Asn-616, Arg-638, and Asp-727 in the editing domaininteract with Asp-402, Glu-410, and Arg-367, respectively, inMid1.

It is remarkable that the position of the editing domainrelative to the aminoacylation domain differs from those inother class II aaRSs with reported structures. For example,compared with E. coli threonyl-tRNA synthetase (ThrRS)(20), the editing domain resides on the opposite side of theaminoacylation domain in AlaRS-�C (Fig. 3A). The amino-acylation active site is �37 Å away from the editing active sitein A. fulgidus AlaRS, which is comparable with the �39 Ådistance in ThrRS (20, 29). We previously obtained a 3.7-Ådataset from an AlaRS-�C crystal belonging to a differentspace group (23). The structure was solved by molecularreplacement, using the present AlaRS-�C structure. Theposition of the editing domain relative to the aminoacylationdomain and their interface are the same.

In most cases, the dimerization of class II aaRSs is mediatedthrough motif 1 (30, 31). Nevertheless, A. aeolicus AlaRS-Nreportedly does not form a dimer, because the tRNA-recognition domain hinders dimerization through motif 1 (14).Consistent with this finding, A. fulgidus AlaRS-�C motif 1 doesnot mediate dimerization, but interacts with the helix–loopstructure (residues 668–688) in the editing domain to form an

ThrRS editing domain AlaRS-∆C editing domain

A

Ile670

Val112

Tyr84

Trp90

Phe678Val685

Tyr681

Arg89

Arg679

Phe107

Trp619

B

Arg367

Asp727

Glu410

Arg638

Phe637

Ile411Leu414

Val356

Thr635

His617Asn616

Asp402

Val403

Thr407

C

Fig. 3. The editing domain. (A) The position of the AlaRS editing domain.The structure of AlaRS-�C, depicted as a tube model, is superposed on that ofThrRS by the aminoacylation domain. The AlaRS-�C model is colored as in Fig.1B, and ThrRS is colored gray. (B) The interface of the editing and aminoacy-lation domains. The helix–loop in the editing domain and motif 1 in theaminoacylation domain are colored salmon and brown, respectively. Theresidues involved in the interactions are shown as white stick models. (C)The interface of the editing and tRNA-recognition domains. The editing coresubdomain and Mid1 are colored orange and blue, respectively.

Naganuma et al. PNAS � May 26, 2009 � vol. 106 � no. 21 � 8491

BIO

CHEM

ISTR

Y

Dow

nloa

ded

by g

uest

on

June

15,

202

1

http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=SF3http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=SF1http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=SF4http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=SF4

interdomain interface. It is interesting to note that AlaX-M,which lacks the helix–loop structure, exists as a monomer insolution (26). In contrast, the helix–loop mediates the ho-modimerization of AlaX-S (27).

The C-Terminal Dimerization Domain. A. fulgidus AlaRS forms adimer through an interaction between the C-terminal dimeriza-tion domains of the 2 molecules (23). AlaRS-�C, lacking thedimerization domain, exists as a monomer in solution (23). Wefirst determined the structure of the isolated dimerizationdomain of A. fulgidus AlaRS, AlaRS-C (Fig. 1C). The structurerevealed that the dimerization domain consists of a long helicalsubdomain and a globular subdomain. The helical subdomaincontains 2 �-helices of 32 and 53 Å and a linker in between. Thissubdomain exclusively mediates the dimer interaction to form acharacteristic HLHZ (Fig. 4A and Fig. S5). Val-744, Met-747,Leu-750, and Leu-751 in �20, and Leu-765, Val-769, Phe-772,Phe-773, Trp-776, Gln-779, Ile-783, Leu-786, Val-789, Ile-790,Leu-793, and Ile-797 in �21, in 1 protomer of the dimer,respectively, form leucine-zipper-like interactions with theircounterparts in the other monomer. The N-terminal portion of

�21 (Pro-762, Leu-765, Pro-766, and Val-769) in 1 protomer alsointeracts with the C-terminal portion of �20 (Ala-754, Ile-757,and Leu-758). The linker mediates the formation of the hydro-phobic core at the �20–�21 junction, where the coiled-coil istwisted (Fig. 4A). Similar HLHZ structures are also present inseveral transcriptional regulators, such as the Myc protoonco-gene product and its relatives (32).

The C-terminal globular subdomain is composed of a6-stranded antiparallel �-sheet and 3 �-helices (Fig. 4B and Fig.S1B). This subdomain tightly packs against �21 of the HLHZ toform a hydrophobic core. Trp-794, Leu-798, and Met-799, in�21, and Val-811, Val-815, Leu-829, Leu-838, and Phe-851, inthe globular subdomain, are involved in the hydrophobic inter-actions. One surface of the globular subdomain is positivelycharged, by the contributions of Lys-855, Arg-859, Arg-863,Arg-867, Lys-870, Arg-876, and Lys-877. It is interesting that theconserved glycine-rich segment (870KGSGGGR876) forms a�-strand that is part of the �-sheet (Fig. 4B). A structuralsimilarity search using the DALI server (33) revealed that theglobular subdomain is similar to that of the ssDNA 5�-3�exonuclease RecJ (34) and exopolyphosphatase (35), with highZ scores of 11.5 and 9.1, respectively.

DiscussiontRNA Interactions. In the crystal structures of tRNA-boundclass-II aaRSs, including ThrRS, seryl-tRNA synthetase, andaspartyl-tRNA synthetase, the amino acid acceptor arm of thetRNA binds to a common site on the class II aminoacylationdomain (20, 36, 37). The binding site corresponds to the grooveformed between Mid1 and Mid2 in the tRNA-recognitiondomain of A. fulgidus AlaRS. However, if the tRNA acceptorarm binds to the groove of A. fulgidus AlaRS via the commonmode (mode 1), then the nucleotides at positions 1–5 and 68–73have a serious steric clash with the archaea-specific insertion ofa helix–loop–helix (InsA2) within the groove (Fig. 5). To avoidthe putative clash, a drastic conformational change should occur.Because InsA2 interacts with �7 and �10 to form a hydrophobiccore, the reorientation of InsA2 seems to be unlikely. When thetRNA relocates to the other tRNA-binding site for editing, it

N N

C C

A

α

α α

B

α

Fig. 4. The C-terminal dimerization domain. (A) The dimer interactions viathe helical subdomains. An HLHZ formed in the dimer is shown in a stereoview.One molecule is colored midnight blue, and the side chains are shown as redstick models. The other molecule is colored gray. (B) The globular subdomainis shown in a ribbon representation. Basic amino acid residues forming thebasic patch are shown as stick models. The conserved Gly-rich segment(870KGSGGGR876) is highlighted in red.

Arg371

Arg731

Val460

Ala-SA

Mid2

InsA2

1 2

Fig. 5. Models of tRNA binding. A tRNA model was created by superpositionof the aminoacylation domains of AlaRS-�C and the E. coli ThrRS�tRNAThr

complex, and the acceptor-arm portion of the tRNA (residues 1–7 and 66–76)is shown as a dark-yellow transparent surface model (mode 1). The secondtRNA model, bound to an alternative tRNA-binding site, is also depicted as ablue surface model (mode 2). The A76 residues in the first and second modelsare highlighted in yellow and blue, respectively. Amino acid residues impor-tant for the aminoacylation or editing activity (17, 28, 38, 39, 45, 46) are shownas stick models. Ala-SA in the aminoacylation site and the editing-site zinc ionare depicted as cpk models. A stereo version of this figure is presented asFig. S6.

8492 � www.pnas.org�cgi�doi�10.1073�pnas.0901572106 Naganuma et al.

Dow

nloa

ded

by g

uest

on

June

15,

202

1

http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=SF5http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=SF1http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=SF1http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=SF6

should dissociate to move over the protuberant Mid2 subdomainand then rebind. However, an alternative mode (mode 2) is thatthe tRNA acceptor stem binds to a groove formed between theMid2 subdomain and the editing domain (‘‘alternative groove’’)of A. fulgidus AlaRS (Fig. 5). In the present structure, the linkerconnecting Mid2 and the editing domain is located in thealternative groove, but the linker appears to be quite flexible andto change its conformation upon tRNA binding. The alternativegroove has entrances to both the aminoacylation and editingactive sites, which would facilitate tRNA relocation betweenthem. Consequently, the alternative mode 2 is more likely thanthe common mode 1 for A. fulgidus AlaRS.

For the bacterial AlaRS from A. aeolicus, the tRNA was dockedin mode 1 (14). Because A.aeolicus AlaRS lacks the archaea-specificinsertion (InsA2), tRNA binding is not hindered. In addition, theacceptor stem could be proximal to Asp-398, corresponding toAla-409 in E. coli AlaRS, which is thought to be indirectly involvedin the G3�U70 interaction (38). The aminoacylation and tRNA-recognition domains of A.aeolicus AlaRS were successfully sepa-rated from the editing domain for crystallography (14), whereas thecorresponding aminoacylation/tRNA-recognition fragment andthe editing domain fragment of E. coli AlaRS were both preparedfor functional studies (17, 38). In contrast, in the case of the archaealAlaRS from A. fulgidus, it was difficult to prepare the correspondingfragments, probably because of the hydrophobic interaction be-tween the aminoacylation/tRNA-recognition domains and the ed-iting domain (Fig. 3). Therefore, the bacterial AlaRSs might havethe editing domain in a different location from that in the archaealAlaRS, relative to the aminoacylation/tRNA-recognition domains,thus allowing the tRNA to shift easily between the 2 active sites. InE. coli AlaRS, the aminoacylation/tRNA-recognition domains andthe editing domain are both able to recognize the G3�U70 base pairin the tRNA acceptor stem (17), involving Arg-314 on the tRNA-recognition domain and Arg-693 on the editing domain of E. coliAlaRS (17, 39). Intriguingly, in the present A. fulgidus AlaRSstructure, Arg-371 (Mid1) and Arg-731, which correspond to theG3�U70 recognition residues Arg-314 and Arg-693, respectively, areclose to the putative tRNA binding site in the alternative groove(Fig. 5, mode 2). The tRNA-recognition domain including Arg-371resides on the minor groove side of the acceptor stem, and thisbinding mode (mode 2) is more preferable for the minor grooverecognition of the G3�U70 base pair than mode 1 (12, 13). There-fore, we cannot exclude the possibility that the bacterial AlaRSshave a similar domain arrangement to that of A. fulgidus AlaRS, andbind tRNA via mode 2. However, the editing domain could interactwith the G3�U70 base pair in the major groove (Fig. 5). Arg-371 andArg-731 are too far from each other to simultaneously interact withthe G3�U70 base pair. Our model in mode 2 is compatible with thefact that E. coli AlaRS aminoacylates the 3�-OH of A76 (40).

The C-terminal dimerization domain of A. fulgidus AlaRS isalso crucial for the tRNA interaction, because the deletion of thedomain dramatically reduces the aminoacylation activity (23).The A. fulgidus AlaRS-C structure reveals that the globularsubdomain of the dimerization domain possesses a positively-charged face, including the conserved Gly-rich segment (Fig.4B). The globular subdomain, therefore, is a candidate for thetRNA-binding site, to support aminoacylation reactions. For E.coli AlaRS, the region of residues 808–875 is a nonspecifictRNA-binding site (17), which includes the Gly-rich segment.

The Full-Length AlaRS Structure. The present structures ofAlaRS-�C and AlaRS-C provide clues to infer the full-lengthAlaRS structure. The distance between the N termini in theAlaRS-C dimer is only 14 Å, which could restrict the positionsof the other domains. Because the editing domain C terminus isconnected to the dimerization domain N terminus, the 2 editingdomains in the dimer should be close to each other, whereas theaminoacylation and tRNA-recognition domains would be dis-

tant from the 2-fold axis. In the AlaRS-�C crystal structure, theediting domain interacts back-to-back with that of the symmetry-related molecule correlated by the crystallographic 2-fold axis.Met-650, Ile-656, Leu-657, and Met-716 mediate the interaction,and the buried surface area is �400 Å2. In the crystal lattice, thedistance between the editing-domain C-termini is �19 Å, whichallows their connection to the dimerization-domain N terminiwithout a large conformation change. Overall, the full-lengthAlaRS dimer is likely to assume a butterfly-like structure (Fig.6). Although this model still requires validation, it could serve asa platform for future analyses.

Materials and MethodsProtein Preparation. See SI Text.

Crystallization and Data Collection. See SI Text.

Structure Determination and Refinement. The structure of AlaRS-�C was solvedby the single-wavelength anomalous dispersion method. The Se-site andinitial phase determinations and solvent flattening were performed with theAutoSHARP program (41). All 15 of the Se sites were identified. Densitymodification and initial model building using the RESOLVE program placed51% of the amino acid residues, and the remaining residues were builtmanually with the COOT program (42, 43). Structure refinement was carriedout with the CNS program (44). A randomly-chosen 5% of the data were setaside for cross-validation. The refinement included several rounds of simulat-ed-annealing, positional, and individual B factor refinements. The refinementconverged to final R and Rfree factors of 21.5% and 26.4%, respectively (TableS1). In the Ramachandran plot, 87.4%, 11.9%, and 0.6% of the residues fell inthe most favored, additional allowed, and generously allowed regions, re-spectively. No residues were in the disallowed region.

The structure of AlaRS-C was solved by the SAD method with the AutoSHARPprogram (41). Of the 56 Se sites, 48 were identified. Model building was per-formed manually by using the COOT program (42). Refinement was done withthe CNS program, and the R and Rfree factors for the final model are 20.5% and27.6%, respectively (Table S1). In the Ramachandran plot, 88.5%, 11.2% and0.2% of the residues fell in the most favored, additional allowed, and generouslyallowed regions, respectively. No residues were in the disallowed region.

ACKNOWLEDGMENTS. We thank the staffs of the Photon Factory (Tsukuba,Japan) and SPring-8 BL41XU (Hyogo, Japan) beam lines for assistance with ourdata collection. This work was supported in part by a Ministry of Education,Culture, Sports, Science, and Technology Global Centers of Excellence Program(Integrative Life Science Based on the Study of Biosignaling Mechanisms), aMinistry of Education, Culture, Sports, Science, and Technology Grant-in-Aid forScientific Research, and the Ministry of Education, Culture, Sports, Science andTechnology Targeted Proteins Research Program. R.F. was supported by ResearchFellowships from the Japan Society for the Promotion of Science.

Fig. 6. A model of the full-length AlaRS dimer. Two copies of AlaRS-�C,which are correlated by the crystallographic 2-fold axis, and an AlaRS-C dimer,are shown. The N termini of AlaRS-C were placed near the C termini ofAlaRS-�C. The model was colored as in Fig. 1.

Naganuma et al. PNAS � May 26, 2009 � vol. 106 � no. 21 � 8493

BIO

CHEM

ISTR

Y

Dow

nloa

ded

by g

uest

on

June

15,

202

1

http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=STXThttp://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=STXThttp://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=ST1http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=ST1http://www.pnas.org/cgi/data/0901572106/DCSupplemental/Supplemental_PDF#nameddest=ST1

1. Jasin M, Regan L, Schimmel P (1983) Modular arrangement of functional domainsalong the sequence of an aminoacyl tRNA synthetase. Nature 306:441–447.

2. Beebe K, Ribas De Pouplana L, Schimmel P (2003) Elucidation of tRNA-dependentediting by a class II tRNA synthetase and significance for cell viability. EMBO J 22:668–675.

3. Cusack S (1995) Eleven down and nine to go. Nat Struct Biol 2:824–831.4. Eriani G, Delarue M, Poch O, Gangloff J, Moras D (1990) Partition of tRNA synthetases

into two classes based on mutually exclusive sets of sequence motifs. Nature 347:203–206.

5. Jasin M, Regan L, Schimmel P (1984) Dispensable pieces of an aminoacyl tRNA syn-thetase which activate the catalytic site. Cell 36:1089–1095.

6. Putney SD, et al. (1981) Primary structure of a large aminoacyl-tRNA synthetase. Science213:1497–1501.

7. Hou YM, Schimmel P (1988) A simple structural feature is a major determinant of theidentity of a transfer RNA. Nature 333:140–145.

8. McClain WH, Foss K (1988) Changing the identity of a tRNA by introducing a G-Uwobble pair near the 3� acceptor end. Science 240:793–796.

9. Vasil’eva IA, Moor NA (2007) Interaction of aminoacyl-tRNA synthetases with tRNA:General principles and distinguishing characteristics of the high-molecular-weightsubstrate recognition. Biochemistry (Moscow) 72:247–263.

10. Beuning PJ, Musier-Forsyth K (1999) Transfer RNA recognition by aminoacyl-tRNAsynthetases. Biopolymers 52:1–28.

11. Francklyn C, Schimmel P (1989) Aminoacylation of RNA minihelices with alanine.Nature 337:478–481.

12. Musier-Forsyth K, Schimmel P (1992) Functional contacts of a transfer RNA synthetasewith 2�-hydroxyl groups in the RNA minor groove. Nature 357:513–515.

13. Musier-Forsyth K, et al. (1991) Specificity for aminoacylation of an RNA helix: Anunpaired, exocyclic amino group in the minor groove. Science 253:784–786.

14. Swairjo MA, et al. (2004) Alanyl-tRNA synthetase crystal structure and design foracceptor-stem recognition. Mol Cell 13:829–841.

15. Swairjo MA, Schimmel PR (2005) Breaking sieve for steric exclusion of a noncognateamino acid from active site of a tRNA synthetase. Proc Natl Acad Sci USA 102:988–993.

16. Lee JW, et al. (2006) Editing-defective tRNA synthetase causes protein misfolding andneurodegeneration. Nature 443:50–55.

17. Beebe K, Mock M, Merriman E, Schimmel P (2008) Distinct domains of tRNA synthetaserecognize the same base pair. Nature 451:90–93.

18. Fukai S, et al. (2000) Structural basis for double-sieve discrimination of L-valine fromL-isoleucine and L-threonine by the complex of tRNAVal and valyl-tRNA synthetase. Cell103:793–803.

19. Fukunaga R, Yokoyama S (2005) Aminoacylation complex structures of leucyl-tRNAsynthetase and tRNALeu reveal two modes of discriminator-base recognition. NatStruct Mol Biol 12:915–922.

20. Sankaranarayanan R, et al. (1999) The structure of threonyl-tRNA synthetase-tRNAThr

complex enlightens its repressor activity and reveals an essential zinc ion in the activesite. Cell 97:371–381.

21. Silvian LF, Wang J, Steitz TA (1999) Insights into editing from an Ile-tRNA synthetasestructure with tRNAIle and mupirocin. Science 285:1074–1077.

22. Tukalo M, Yaremchuk A, Fukunaga R, Yokoyama S, Cusack S (2005) The crystal structureof leucyl-tRNA synthetase complexed with tRNALeu in the post-transfer-editing con-formation. Nat Struct Mol Biol 12:923–930.

23. Fukunaga R, Yokoyama S (2007) Crystallization and preliminary X-ray crystallographicstudy of alanyl-tRNA synthetase from the archaeon Archaeoglobus fulgidus. ActaCrystallogr F 63:224–228.

24. Ahel I, Korencic D, Ibba M, Söll D (2003) Trans-editing of mischarged tRNAs. Proc NatlAcad Sci USA 100:15422–15427.

25. Schimmel P, Ribas De Pouplana L (2000) Footprints of aminoacyl-tRNA synthetases areeverywhere. Trends Biochem Sci 25:207–209.

26. Fukunaga R, Yokoyama S (2007) Structure of the AlaX-M trans-editing enzyme fromPyrococcus horikoshii. Acta Crystallogr D 63:390–400.

27. Sokabe M, Okada A, Yao M, Nakashima T, Tanaka I (2005) Molecular basis of alaninediscrimination in editing site. Proc Natl Acad Sci USA 102:11669–11674.

28. Hill K, Schimmel P (1989) Evidence that the 3� end of a tRNA binds to a site in theadenylate synthesis domain of an aminoacyl-tRNA synthetase. Biochemistry 28:2577–2586.

29. Dock-Bregeon AC, et al. (2004) Achieving error-free translation; the mechanism ofproofreading of threonyl-tRNA synthetase at atomic resolution. Mol Cell 16:375–386.

30. Logan DT, Mazauric MH, Kern D, Moras D (1995) Crystal structure of glycyl-tRNAsynthetase from Thermus thermophilus. EMBO J 14:4156–4167.

31. Mosyak L, Reshetnikova L, Goldgur Y, Delarue M, Safro MG (1995) Structure ofphenylalanyl-tRNA synthetase from Thermus thermophilus. Nat Struct Biol 2:537–547.

32. Nair SK, Burley SK (2003) X-ray structures of Myc-Max and Mad-Max recognizing DNA.Molecular bases of regulation by protooncogenic transcription factors. Cell 112:193–205.

33. Holm L, Sander C (1998) Touring protein fold space with Dali/FSSP. Nucleic Acids Res26:316–319.

34. Yamagata A, Kakuta Y, Masui R, Fukuyama K (2002) The crystal structure of exonu-clease RecJ bound to Mn2� ion suggests how its characteristic motifs are involved inexonuclease activity. Proc Natl Acad Sci USA 99:5908–5912.

35. Ugochukwu E, Lovering AL, Mather OC, Young TW, White SA (2007) The crystalstructure of the cytosolic exopolyphosphatase from Saccharomyces cerevisiae revealsthe basis for substrate specificity. J Mol Biol 371:1007–1021.

36. Cavarelli J, et al. (1994) The active site of yeast aspartyl-tRNA synthetase: Structural andfunctional aspects of the aminoacylation reaction. EMBO J 13:327–337.

37. Biou V, Yaremchuk A, Tukalo M, Cusack S (1994) The 2.9-Å crystal structure of T.thermophilus seryl-tRNA synthetase complexed with tRNASer. Science 263:1404–1410.

38. Ho C, Jasin M, Schimmel P (1985) Amino acid replacements that compensate for a largepolypeptide deletion in an enzyme. Science 229:389–393.

39. Ribas de Pouplana L, Buechter D, Sardesai NY, Schimmel P (1998) Functional analysis ofpeptide motif for RNA microhelix binding suggests new family of RNA-binding do-mains. EMBO J 17:5449–5457.

40. Hecht SM, Chinualt AC (1976) Position of aminoacylation of individual Escherichia coliand yeast tRNAs. Proc Natl Acad Sci USA 73:405–409.

41. Emsley P, Cowtan K (2004) Coot: Model-building tools for molecular graphics. ActaCrystallogr D 60:2126–2132.

42. Vonrhein C, Blanc E, Roversi P, Bricogne G (2006) Automated structure solution withautoSHARP. Methods Mol Biol 364:215–230.

43. Terwilliger TC (2000) Maximum-likelihood density modification. Acta Crystallogr D56:965–972.

44. Brunger AT, et al. (1998) Crystallography and NMR system: A new software suite formacromolecular structure determination. Acta Crystallogr D 54:905–921.

45. Davis MW, Buechter DD, Schimmel P (1994) Functional dissection of a predictedclass-defining motif in a class II tRNA synthetase of unknown structure. Biochemistry33:9904–9911.

46. Shi JP, Musier-Forsyth K, Schimmel P (1994) Region of a conserved sequence motif in aclass II tRNA synthetase needed for transfer of an activated amino acid to an RNAsubstrate. Biochemistry 33:5312–5318.

8494 � www.pnas.org�cgi�doi�10.1073�pnas.0901572106 Naganuma et al.

Dow

nloa

ded

by g

uest

on

June

15,

202

1