Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Engineering a Selectable Marker for Hyperthermophiles *

Engineering a Selectable Marker for Hyperthermophiles * THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 280, No. 12, Issue of March 25, pp. 11422–11431, 2005 © 2005 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. □ S Received for publication, December 3, 2004, and in revised form, December 27, 2004 Published, JBC Papers in Press, January 7, 2005, DOI 10.1074/jbc.M413623200 Stan J. J. Brouns‡§, Hao Wu‡, Jasper Akerboom‡, Andrew P. Turnbull , Willem M. de Vos‡, and John van der Oost‡ From the ‡Laboratory of Microbiology, Department of Agrotechnology and Food Sciences, Wageningen University, Hesselink van Suchtelenweg 4, 6703 CT Wageningen, The Netherlands and Proteinstrukturfabrik c/o Berliner Elecktronenspeicherring-Gesellschaft fu¨r Synchrotronstrahlung GmbH, Albert Einstein Strasse 15, D-12489 Berlin, Germany above 80 °C, improving the thermal stability of a protein is still Limited thermostability of antibiotic resistance mark- ers has restricted genetic research in the field of ex- a challenging task. This is mainly because the laws governing tremely thermophilic Archaea and bacteria. In this protein stability are not easily extracted because they are study, we used directed evolution and selection in the highly variable and complex (1, 2). It seems generally accepted thermophilic bacterium Thermus thermophilus HB27 to that the extreme stability of certain natural proteins results find thermostable variants of a bleomycin-binding pro- from the cumulative effect of small adaptations in protein tein from the mesophilic bacterium Streptoalloteichus architecture and amino acid composition. Although some of hindustanus. In a single selection round, we identified these stabilizing features, such as optimized surface ion pair eight clones bearing five types of double mutated genes networks (3), are unlikely to be engineered into a protein of that provided T. thermophilus transformants with bleo- interest, other strategies like -helix capping (4) and the intro- mycin resistance at 77 °C, while the wild-type gene could duction of disulfide bonds and prolines in -turns (5) can be only do so up to 65 °C. Only six different amino acid applied very successfully when carefully designed on the basis positions were altered, three of which were glycine res- of a high resolution crystal structure. However, in many cases idues. All variant proteins were produced in Esche- atomic resolution three-dimensional information of a protein is richia coli and analyzed biochemically for thermal sta- unavailable. Directed evolution approaches, by contrast, do not bility and functionality at high temperature. A synthetic require any structural information and commonly rely on ran- mutant resistance gene with low GC content was de- signed that combined four substitutions. The encoded dom mutagenesis and recombination followed by screening or protein showed up to 17 °C increased thermostability selection schemes (1, 6). Thermostability screens of mutant and unfolded at 85 °C in the absence of bleomycin, libraries are usually carried out by applying a thermal chal- whereas in its presence the protein unfolded at 100 °C. lenge at nonpermissive temperatures after which the remain- Despite these highly thermophilic properties, this mu- ing functionality of the individual clones is tested (7, 8). To tant was still able to function normally at mesophilic explore sufficient sequence space requires the testing of large temperatures in vivo. The mutant protein was co-crys- numbers of mutant clones, which necessitates high throughput tallized with bleomycin, and the structure of the binary approaches such as the use of robotics. Conversely efficient complex was determined to a resolution of 1.5 Å. De- selection procedures allow the testing of a large set of variants tailed structural analysis revealed possible molecular while reducing the effort of finding improved ones to mechanisms of thermostabilization and enhanced anti- a minimum. biotic binding, which included the introduction of an A convenient selection system for finding protein variants in intersubunit hydrogen bond network, improved hydro- a library with improved thermostability is based on in vivo phobic packing of surface indentations, reduction of screening in a thermophilic expression host. Cloning and selec- loop flexibility, and -helix stabilization. The potential tion in thermophilic microorganisms such as Geobacillus applicability of the thermostable selection marker is discussed. stearothermophilus (30–60 °C) or Thermus thermophilus (50– 80 °C) mimic natural evolution but are only applicable when the gene of interest encodes a protein that is of biological Despite the vast amount of protein sequences and structures relevance to growth or survival of the host organism (9, 10). from microorganisms that grow optimally at temperatures The selective pressure can be fine tuned by raising the temper- ature of growth, enabling only hosts that bear thermoadapted * This work was supported by a grant from the European Union in variants to grow on solid media. For instance, a combination of the framework of the SCREEN project (Contract QLK3-CT-2000- in vitro mutagenesis methods and in vivo selection schemes 00649). The costs of publication of this article were defrayed in part by have led to a highly thermostable kanamycin nucleotidyltrans- the payment of page charges. This article must therefore be hereby ferase gene that is able to function at temperatures up to 79 °C marked “advertisement” in accordance with 18 U.S.C. Section 1734 (11). Such mutant selection markers have permitted the devel- solely to indicate this fact. □ S The on-line version of this article (available at http://www.jbc.org) opment of genetic tools that are very useful in the study of contains Supplemental Table 1. gene-function relationships in thermophilic bacteria (12). The nucleotide sequence(s) reported in this paper has been submitted TM In contrast to thermophiles (optimum temperature for to the GenBank /EBI Data Bank with accession number(s) AY780486. The atomic coordinates and structure factors (code 1XRK) have been growth, 60–80 °C), antibiotic-based genetic systems for hyper- deposited in the Protein Data Bank, Research Collaboratory for Struc- thermophilic bacteria and Archaea (optimum temperature for tural Bioinformatics, Rutgers University, New Brunswick, NJ growth, 80 °C) are still in their infancy. This is primarily due (http://www.rcsb.org/). to the absence of thermostable antibiotics and their corre- § To whom correspondence should be addressed. Tel.: 31-317-483110; Fax: 31-317-483829; E-mail: [email protected]. sponding resistance factors because most known antibiotic- 11422 This paper is available on line at http://www.jbc.org This is an Open Access article under the CC BY license. Crystal Structure of Thermostable Bleomycin-binding Protein 11423 kanamycin nucleotidyltransferase gene. Ligation mixtures were trans- producing microorganisms are mesophilic bacteria and fungi. formed into E. coli HB101, and transformants were plated on 1.5% LB Often the common antibiotics cannot be used because many of agar plates supplemented with 3 g/ml bleomycin. Both blmS and shble them are unstable at high temperatures, or hyperthermophiles provided resistance against the antibiotic, giving rise to the 4434-bp are simply insensitive to them (13). The glycopeptide bleomycin plasmid pWUR111 and the 4404-bp plasmid pWUR112, respectively. is an exception since it is a highly thermostable molecule and Mutant Library Construction—Error-prone PCR was carried out us- effective against many aerobic microorganisms and eukaryotic ing two different polymerases, namely Taq (Amersham Biosciences) and Mutazyme (Genemorph kit, Stratagene). This approach was chosen cell lines (14, 15). The bleomycin family of antibiotics, including to complement the transition and transversion bias of each enzyme to phleomycin and tallysomycin, are DNA- and RNA-cleaving gly- provide a more complete mutational spectrum of the PCR product. For copeptides that are produced by the actinomycetes Strep- the error-prone amplification, flanking primers BG1412 (sense, 5- toalloteichus hindustanus and Streptomyces verticillus. As lit- CGACCCTTAAGGAGGTGTGAGGCATATG-3) and BG1408 (anti- tle as a few hundred bleomycin molecules can effectively kill sense, 5-CGAGCTCGGTACCCGGGGATCCTCTAGATTA-3) (NdeI aerobic cells (16). For this reason, bleomycin is currently clin- and XbaI sites are underlined) were designed to allow variation throughout the entire coding sequence between the start and stop codon ically used as an antitumor agent against squamous cell carci- (indicated in boldface). Taq polymerase-based PCRs were performed as nomas and malignant lymphomas (17). Resistance against described previously (29). A 50-l PCR contained 5 ng of pWUR112, 5 bleomycin-like antibiotics is conferred by N-acetylation, deami- pmol of each primer, 0.2 mM dATP and dGTP, 1 mM dCTP and dTTP, 5 dation, and sequestration of the molecule (15). The latter mech- units of polymerase, 3 mM MgCl , and three concentrations of MnCl 2 2 anism involves bleomycin-binding proteins (BBPs), which (0.1, 0.3, and 0.5 mM). The mixture was thermocycled as follows: 95 °C have been found only in mesophilic bacteria. Two proteins, (4 min); 30 cycles of 94 °C (30 s), 55 °C (45 s), and 72 °C (25 s); and postdwelled for 4 min at 72 °C. Mutazyme PCRs were prepared accord- Shble and BlmA, provide self-immunity for bleomycin produc- ing to the manufacturer’s instructions and thermocycled as above using ers S. hindustanus and S. verticillus, respectively (14, 18), and an elongation time of 50 s. may be involved the transport and excretion of the molecule Randomly mutated PCR products were cloned into vector pMK18 (19). Two genes, blmT and blmS, are located on the Klebsiella and transformed into E. coli HB101. A total of 10,000 Taq- and 10,000 pneumoniae transposon Tn5 (20) and on the Staphylococcus Mutazyme-derived clones were resuspended in 50 ml of LB medium aureus plasmid pUB110 (21), respectively. All four proteins are supplemented with antibiotic and grown in 1 liter of medium to early stationary phase. Plasmids were subsequently harvested using a Mini- highly negatively charged cytoplasmic proteins of around 14 prep plasmid isolation kit (Qiagen). kDa that form homodimers that bind two positively charged Selection in T. thermophilus—T. thermophilus HB27 was kindly antibiotic molecules at a hydrophobic subunit interface cleft provided by Dr. J. Berenguer (Autonomous University of Madrid, Ma- (19, 22, 23). The small protein size and the wide applicability of 2 drid, Spain). Cells were routinely cultivated at 70 °C in a Ca - (3.9 mM) the drug have made both shble and blmT popular dominant and Mg (1.9 mM)-rich medium (28) containing 8 g/liter tryptone, 4 selection markers in vector systems for lower and higher eu- g/liter yeast extract, and 3 g/liter NaCl dissolved in Evian mineral water (pH 7.7, after autoclaving) (Evian-les-Bains, France). Transfor- karyotes, bacteria, and halophilic Archaea (15, 24, 25). This mation of T. thermophilus was essentially performed by the method of prompted us to investigate whether we could thermostabilize Koyama (30). Frozen cell aliquots were resuspended in 25 ml of medium Shble and BlmS to allow for their application in aerobic ther- and grown at 150 rpm to an A of 0.8. The culture was then diluted 1:1 mophiles and hyperthermophiles. in preheated medium and incubated for another hour. Next plasmids In this study, we performed directed evolution using selec- were added to 0.5 ml of culture, and the mixture was incubated for 2–3 tion in the thermophilic bacterium T. thermophilus and ob- h at 70 °C with occasional shaking before being plated on 3% agar plates (BD Biosciences) supplemented with 30 g/ml kanamycin or 15 tained various mutant proteins that could operate under highly g/ml bleomycin (Calbiochem) for selection. Colonies appeared within thermophilic growth conditions. Their enhanced performance 36 h at 60–70 °C. At temperatures above 70 °C, 1% Gelrite plates (Roth, at high temperature was analyzed biochemically, and possible Karlsruhe, Germany) were used supplemented with 100 g/ml kana- stabilizing effects were identified. mycin or 20 g/ml bleomycin for selection. Colonies were grown over- night in liquid medium containing 30 g/ml kanamycin or 5 g/ml EXPERIMENTAL PROCEDURES bleomycin. T. thermophilus plasmid DNA was prepared using a plasmid All chemicals were of analytical grade and purchased from Sigma. Miniprep kit (Qiagen) after a preincubation with 2 mg/ml lysozyme for Primers were obtained from MWG Biotech AG (Ebersberg, Germany). 30 min at 37 °C. Polymerase chain reactions were performed with Pfu TURBO (Strat- Gene Cloning, Overexpression, and Protein Purification—Wild-type agene) unless stated otherwise. Bleomycin A2 (Bleocin, Calbiochem) and double mutant shble genes were PCR-amplified from their respec- was used for all selections. Escherichia coli HB101 (F hsdS20 (r , tive pWUR112 plasmids using primers BG1503 (sense, 5-GATGGC- m ) ara-14 galK2 lacY1 leuB6 mcrB mtl-1 proA2 recA13 rpsL20 CATGGCCAAGTTGACCAGTGC-3) and BG1504 (antisense, 5-GCCG- supE44 thi-1 xyl-5 (Str )) (26) was used for cloning purposes and rou- CAAGCTTAGTCCTGCTCCTCGGCC-3) (NcoI and HindIII sites are tinely transformed by electroporation. underlined). PCR products were cloned into vector pET26b (Novagen) Generation of a Bleomycin-based Shuttle Vector—Bacillus subtilis and fused to an Erwinia carotovora pectate lyase (pelB) signal sequence 168 8G5 carrying pUB110 was kindly provided by Dr. S. Bron (Univer- allowing periplasmic protein overexpression in E. coli BL21(DE3) (No- sity of Groningen, Groningen, The Netherlands), and the plasmid was vagen). Periplasmic fractions of 1-liter cultures were prepared by os- isolated by Qiagen Miniprep according to the manufacturer’s instruc- motic shock according to the manufacturer’s instructions and dialyzed tions. The blmS gene was PCR-amplified with primers BG1407 (sense, overnight against 20 mM Tris-HCl (pH 7.5). Samples were loaded onto 5-GGAGGTGCATATGAGAATGTTACAGTCTATCCC-3) and BG1240 a MonoQ HR 5/50 column connected to a fast protein liquid chromatog- (antisense, 5-CGCGTCTAGATTAGCTTTTTATTTGTTGAAAAAAG- raphy system (Amersham Biosciences) and eluted using a 1 M NaCl 3) (NdeI and XbaI sites are underlined). Chromosomal DNA of S. gradient. Shble-containing fractions were pooled and dialyzed against a hindustanus (ATCC 31158) was prepared according to standard proce- 10 mM NaP buffer (pH 7.0) supplemented with 50 mM NaCl and dures (27) and used for PCR amplification of the shble gene with subsequently purified by size exclusion chromatography using a Super- primers BG1410 (sense, 5-TGAGGCATATGGCCAAGTTGACCAGT- dex 200 HR 10/30 column (Amersham Biosciences). GCCG-3) and BG1411 (antisense, 5-GATCCTCTAGATTAGTCCT- Synthetic Gene Construction—A synthetic mutant shble gene based GCTCCTCGGCCACG-3) (NdeI and XbaI sites are underlined). PCR on archaeal codon usage was constructed by oligonucleotide assembly products were digested and ligated into E. coli-T. thermophilus shuttle PCR (31). This gene contains the point mutations G18E, D32V, L63Q, vector pMK18 (28) (Biotools, Madrid, Spain) thereby replacing the and G98V and has a GC content of 40.8% compared with 70.2% of wild-type shble. The synthetic gene was designated HTS (high temper- TM ature Shble). The sequence has been deposited in GenBank under The abbreviations used are: BBP, bleomycin-binding protein; Shble, accession number AY780486. BBP from S. hindustanus; BlmA, BBP from S. verticillus; BlmT, BBP Assembly PCR mixtures contained 10 oligonucleotides (BG1542– from transposon Tn5; BlmS, BBP from plasmid pUB110; HTS, high BG1451, Supplemental Table 1) with an overlap of 20 bases. Both temperature Shble; r.m.s.d., root mean square deviation; DSC, differ- ential scanning calorimetry. flanking primers were 40 bases in length, whereas the eight central 11424 Crystal Structure of Thermostable Bleomycin-binding Protein primers consisted of 80–90 bases. The assembly PCR mixture contained protein was extensively dialyzed against 10 mM NaP , pH 7.0, and was a 2.5 M concentration of each primer, 0.2 mM dNTPs, and 0.05 units/l subsequently crystallized using the sitting drop method of vapor diffu- Pfu polymerase. The mixture was thermocycled at 94 °C (30 s), 55 °C sion at 20 °C and a protein concentration of 3.3 mg/ml in the presence (30 s), and 72 °C (60 s) for 40 cycles. The PCR products were purified of a 10-fold molar excess of bleomycin A2 HCl (Calbiochem). Crystals over a PCR purification column (Qiagen) and diluted 1:1 in fresh PCR grew optimally using 2.0 M ammonium sulfate as the precipitant in 0.1 mixture containing only both flanking primers BG1542 and BG1551 at M sodium acetate buffer, pH 4.6. Data were collected from a single flash 0.1 M concentration and were thermocycled according to standard frozen native crystal (100 K) to 1.5-Å resolution using a MAR345 procedures. PCR products of the expected size were isolated from an imaging plate at the Protein Structure Factory beamline BL14.2 of the agarose gel using the Qiaex II gel extraction kit (Qiagen), digested with Free University of Berlin at the BESSY synchrotron source (Berlin, NdeI and BglII, and cloned into vector pET26b. This allowed for effi- Germany). All data were reduced with DENZO and SCALEPACK (34). cient cytoplasmic protein overproduction in E. coli BL21(DE3)-RIL The crystal used for data collection had unit cell parameters of a  44.0 (Novagen). Positive clones were picked from LB agar plates containing Å, b  66.6 Å, and c  47.2 Å and   117.4° and belonged to space 3 g/ml bleomycin and 50 g/ml chloramphenicol. A 4-liter culture was group P2 with a dimer in the asymmetric unit. grown at 37 °C to an A of 0.5, induced with 0.5 mM isopropyl--D- The structure of HTS was determined by molecular replacement thiogalactopyranoside, and incubated for another 5 h. Cells were har- using the program MOLREP (35) and the S. verticillus BlmA dimer vested, resuspended in 20 mM Tris-HCl (pH 7.5), and sonicated. Cell (Protein Data Bank code 1JIE) (23) as the search model. The initial extracts were clarified by centrifugation (30 min at 26,500  g at 4 °C) phases were improved using the free atom refinement method together and applied to a 70-ml Q-Sepharose Fast Flow (Amersham Biosciences) with automatic model tracing in ARP/wARP (36). Translation, libra- anion exchange column. Proteins were eluted by a 1 M NaCl gradient, tion, and screw rotation (TLS) parameters were determined, and TLS- and HTS-containing fractions were pooled and concentrated over a restrained refinement was performed using REFMAC (37). Several YM10 filter (Amicon) and further purified by size exclusion chromatog- rounds of iterative model building and refinement followed, and water raphy as described above. molecules were added using ARP/wARP (36). The final model (compris- DNA Sequencing—Inserts of plasmids used in this study were se- ing 241 amino acids, 288 water molecules, two bleomycin molecules, quenced by Westburg Genomics (Wageningen, The Netherlands). and three sulfate ions), refined using data between 30- and 1.5-Å Protein Quantitation—Protein concentrations were determined by resolution, has an R- and free R-factor of 17.4 and 19.4%, respectively, using a Bradford assay (32) (Bio-Rad). Purified proteins were quantified with good geometry. Residues Met-1, Glu-122, Gln-123, and Asp-124 in from A measurements using a protein extinction coefficient of 29,000 chain A and Met-1, Gln-123, and Asp-124 in chain B are not visible in 1 1 M cm (33). the electron density map and therefore have been excluded from the DNA Protection Assay—Protein functionality assays were essentially model. Additionally the side chains of two residues in chain A (Asp-36 performed as described elsewhere (14) using a 10-fold excess molar and Arg-87), four side chains in chain B (Glu-21, Asp-36, Arg-87, and concentration of protein over bleomycin A2 (Calbiochem). In each assay, Glu-122), and the -aminopropyldimethylsulfonium moiety of bleomy- 0.2 g of PstI-linearized plasmid, pUC19, was used. Assays were per- cin have been truncated in the final model. The stereochemical quality formed by first incubating DNA and protein shortly at 85 °C after which of the model and the model fit to the diffraction data were analyzed with bleomycin A2, dithiothreitol, and Fe were sequentially added to the the programs PROCHECK (38) and SFCHECK (39). reaction mixture. The coordinates and experimental structure factors have been depos- Circular Dichroism Spectroscopy—CD experiments were performed ited in the Protein Data Bank under accession code 1XRK. Figures were on a Jasco J-715 spectropolarimeter (Jasco, Tokyo, Japan) equipped prepared with Swiss-PDBviewer version 3.7 SP5 (40) and rendered with a PTC-348WI Peltier temperature control system. Far-UV CD with POV-Ray version 3.6. measurements were conducted with Suprasil quartz cuvettes (Hellma Benelux, Rijswijk, The Netherlands) with a 1-mm cell length. During RESULTS AND DISCUSSION all experiments, the sample cell chamber was purged by dry N gas at Selection of Stabilized Variants of Shble a flow rate of 10 liters/min. In temperature-induced unfolding experi- ments, the cuvette containing a 1.7 M concentration of protein sample Randomly mutated shble genes were introduced in the in degassed 10 mM NaP (pH 7.0) and 50 mM NaCl was heated from 25 i E. coli-T. thermophilus shuttle vector pMK18 under the control to 95 °C at 0.4 °C/min and subsequently cooled to 25 °C at the same of the promoter of the surface layer protein A (slpA) from rate. The ellipticity at 205 nm was measured every 0.5 °C with a 2-s T. thermophilus HB8 (41). This promoter is known to drive response time to monitor the loss of -sheet and -turn secondary efficient transcription of the single selection marker in both structure elements. The bandwidth of the measurement was set to 1.0 bacteria. An error-prone library of 20,000 functional clones nm, and the sensitivity was set to 100 millidegrees. Data were corrected for the temperature-dependent ellipticity of a blank without protein. was generated in E. coli HB101. Colonies appeared of similar Averaged data of two independent scans were fit according to a two- size, and there was no difference between the mutant and state model of unfolding, and the apparent temperature unfolding mid- wild-type shble phenotype. The plasmid library was harvested point (T ) was derived from van ’t Hoff plots. and transformed into T. thermophilus HB27, making use of its Fluorescence Spectroscopy—Fluorescence experiments were per- high natural competence (30). Thermus clones appeared on formed on a Varian Cary Eclipse fluorometer (Varian, Middelburg, The bleomycin-containing plates up to 65 °C after transformation Netherlands) equipped with a four-cuvette Peltier multicell holder and PCB-150 waterbath. All measurements were performed in 3-ml Supra- with the wild-type shble shuttle vector, whereas wild-type sil quartz cuvettes (Hellma Benelux) with a 1-cm path length. A mag- blmS was unable to generate a resistant phenotype at either 50 netic stirring bar ensured a homogeneous sample temperature. The or 65 °C. The transformation efficiency of the shble shuttle temperature of the sample was recorded by a temperature probe inside vector was approximately 5 times lower at 65 °C than the one of the four samples. Spectra and thermal unfolding curves were kanamycin-based vector pMK18 (28). This difference might be recorded of 1.7 M protein solutions in degassed 10 mM NaP (pH 7.0) due to the lethal effect of bleomycin and the non-catalytic buffer supplemented with 50 mM NaCl. Tryptophans were excited at 295 nm, and fluorescence was recorded from 300 to 550 nm with both nature of its elimination, which requires at least one protein the excitation and the emission slits set to 5 nm. During temperature- molecule per bleomycin molecule. induced unfolding and refolding studies, fluorescence emission intensi- Upon increasing the temperature of selection, a dramatic ties were monitored at 315 nm from 27 to 92 °C at a heating and cooling decrease in the number of colonies was observed after trans- rate of 0.4 °C/min. Data were corrected for the fluorescence emission of formation of 8 g of mutant library DNA. While 1200 colonies corresponding blank solutions. Data of two independent scans were appeared at 67 °C, this number decreased to 800 at 69 °C, 600 treated and fit as described above. Differential Scanning Calorimetry (DSC)—DSC measurements were at 70 °C, 106 at 75 °C, and eight at 77 °C. No colonies appeared performed on a Microcal III system (Setaram, Caluire, France). De- at 78 and 80 °C. Plating efficiencies at these temperatures have gassed protein samples of 0.28 mg/ml (20 M)in10mM NaP (pH 7.0) been reported to be severely reduced, complicating selection up and 50 mM NaCl in the presence and absence of an 8-fold molar excess to 85 °C, the maximum temperature of growth (11). The eight of bleomycin A2 were heated from 20 to 120 °C at 0.5 °C/min. Midpoint Thermus clones found at 77 °C (termed 77-1 to 77-8) were temperatures of unfolding were determined by curved base-line analy- grown overnight in selective medium at 70 °C; their plasmids sis from two independent scans. Protein Crystallization, Data Collection, and Processing—The HTS were isolated, transformed into E. coli HB101, and subse- Crystal Structure of Thermostable Bleomycin-binding Protein 11425 TABLE I double mutants, only 77-3 (D32V,L63Q) seems to have a Nucleotide and amino acid substitutions of Shble variants marked increase in T as observed with CD spectroscopy and Residue number fluorescence spectroscopy, while numbers 77-1 (L63Q,G98V), 77-5 (R31L,G98S), and 77-7 (G18E,L63Q) remain virtually un- 18 31 32 40 63 98 changed. Surprisingly mutant 77-4 (R31L,G40A) displays sig- Wild type Gly Arg Asp Gly Leu Gly nificantly lower T values compared with the wild-type. Quad- Gln Val 77-1/77-2/77-8 T188A G293T ruple mutant HTS, which combines non-redundant mutations Val Gln found in 77-1, 77-3, and 77-7, displays a profound increase of 77-3 A95T T188A 13.9, 10.8, and 17.7 °C in stability in the absence of the anti- Leu Ala 77-4/77-6 biotic as found by CD, fluorescence spectroscopy, and DSC, G92T G119C Leu Ser respectively. In its presence, the complex becomes hyperther- 77-5 G92T G292A mostable, unfolding at a temperature of just over 100 °C, 5.6 °C Glu Gln 77-7 higher compared with the wild-type protein. G53A T188A To our surprise, the double mutants 77-1, 77-5, and 77-7 had Clone contained silent nucleotide substitution C360A. almost unchanged apparent melting temperatures compared Clone contained silent nucleotide substitution G63A. c with the wild-type. This can be understood by realizing that in Clone contained silent nucleotide substitution G30A. vivo some amino acid changes may prevent instances of local protein unfolding and therefore may avoid further unfolding quently reisolated; and their inserts were sequenced. This re- and subsequent proteolytic attack. However, this is not neces- vealed that all variants were double mutants bearing, in total, sarily reflected in its in vitro melting temperature, which is a six different amino acid substitutions and three silent muta- measure of its global stability. Only when the weakest point of tions (Table I). Five types of double mutants could be distin- a structure was compensated (D32V and L63Q in 77-3), an guished at the protein level, and two sets of double mutants increase of its melting temperature from 70.8 to 79.5 °C with were identical. Remarkably three of six mutations found were CD and 67.9 to 69.1 °C with fluorescence spectroscopy was glycine substitutions of which glycine 98 was replaced by either observed. Adding mutation G98V from 77-1 and G18E from a valine or a serine. The fact that only double mutants were 77-7 to 77-3, giving rise to HTS, further increased its melting found seems to be a clear indication of the high stringency that temperature as one would expect. This observation is analo- was used during selection. Interestingly some substitutions, gous to the findings of extensive work that has been conducted such as L63Q, had occurred in combination with either G18E, with the neutral protease from G. stearothermophilus where D32V, or G98V, which may point to the independent effects of interactions close to the N terminus were found to be limiting the different mutations. A multiple sequence alignment of the global stability (5). BBPs and the position of the mutations are shown in Fig. 1. To assess the reason why these mutants performed better at ele- Mutants Improve DNA Protection against Bleomycin at vated temperatures in vivo, we produced and purified wild-type High Temperature Shble and all double mutants and studied their biochemical In vitro DNA protection assays were performed with the behavior in vitro. Furthermore a synthetic quadruple mutant gene with low GC content was designed by combining muta- various Shble mutants to test whether the resistant phenotype of T. thermophilus at 77 °C was due to improved protection tions G18E, D32V, L63Q, and G98V. The protein, designated HTS, was produced, purified, and biochemically analyzed. The against the DNA degrading capability of bleomycin. The result of this is shown in Fig. 2. At 25 °C, no significant differences in HTS protein was crystallized in complex with bleomycin A2, and its structure was determined. band intensities were observed. A 30-min thermal preincuba- tion of the protein at 85 °C, however, revealed a drastic loss of Thermal Unfolding function in mutant 77-4. Differences between the wild-type and mutants became pronounced when bleomycin binding capabil- Shble variants were subjected to temperature-induced equi- ities were tested at 85 °C. At this temperature, the DNA was librium unfolding experiments in the presence and absence of protected best by 77-1 and HTS followed by 77-4, 77-5, 77-3, bleomycin. The protein was found to unfold largely irreversibly 77-7, and the wild type. Surprisingly mutant 77-4, which dis- since only 40% of the native folded signal was regained after played a low temperature unfolding midpoint and high thermal slow cooling of the thermally unfolded protein. Therefore only inactivation at 85 °C, apparently bound bleomycin effectively apparent midpoint temperatures of unfolding (T ) could be at high temperature conditions. So although the global stability calculated. The results are summarized in Table II. of this mutant was decreased, it had improved bleomycin bind- In the absence of the antibiotic, wild-type Shble appears to be ing characteristics, which in itself stabilizes the protein dra- a very stable protein. This is remarkable because S. hindusta- matically as observed by DSC measurements for the wild type. nus grows optimally at 28 °C (42). It is often found, however, These results indicate that some of the double mutants have that proteins for which low biological turnover is beneficial for improved the bleomycin binding properties compared with the a host are prone to little local unfolding and hence are less wild type, confirming the findings of the in vivo selection pro- susceptible to proteolytic attack (43). Structurally Shble, which cedure in T. thermophilus. Possible structural explanations for serves a function of self-immunity, might well be adapted to the improved functionality at higher temperature are dis- meet these criteria by its compactness, relatively high second- cussed below. ary structure content, high surface charge, and embedded N and C termini (19). The unfolding data also clearly show the Overall Structure Description strong stabilizing effect of ligand binding on the thermostabil- ity of the BBP since the apparent unfolding midpoint temper- The quadruple mutant HTS was crystallized in the presence ature increases 27.3 °C upon bleomycin binding. This effect has of bleomycin A2, and its structure was determined to 1.50-Å also been recognized in other ligand-binding proteins, such as resolution (Table III). The crystals grown belong to space group streptavidin and avidin, which become extremely thermostable P2 with unit cell parameters of a  44.0 Å, b  66.6 Å, and c in the presence of biotin (44). In the absence of bleomycin, the 47.2 Å and   117.4° and a dimer in the asymmetric unit (Fig. stability of the various mutants is rather different. Of the 3, A and B). Representative electron density is shown in Fig. 11426 Crystal Structure of Thermostable Bleomycin-binding Protein FIG.1. Structural alignment of bleomycin-binding proteins from different microbial sources. Shble (Protein Data Bank code 1BYL) from S. hindustanus (19), BlmA (Protein Data Bank code 1QTO) from S. verticillus ATCC15003 (45), SvP from S. verticillus ATCC21890, BlmT (Protein Data Bank code 1ECS) from K. pneumoniae transposon Tn5 (22), and BlmS from S. aureus plasmid pUB110 are shown. The alignment was created by backbone superimposition of the three structures and expanded with the SvP and BlmS sequences by realignment using ClustalX version 1.81 (65) while maintaining the original gaps. The HTS structure was used for residue numbering and topology assignment (black arrows, -strand; checkered boxes, -helix). Mutations are indicated by arrows. TABLE II Apparent thermal unfolding midpoints (°C) of Shble variants WT HTS 77-1 77-3 77-4 77-5 77-7 Uncomplexed CD 70.8  0.5 84.7  0.6 72.2  0.5 79.5  0.6 66.8  0.5 71.1  0.6 70.3  0.5 FS 67.9  0.4 78.7  0.5 67.2  0.4 69.1  0.5 63.8  0.4 64.6  0.5 65.9  0.4 DSC 67.4  0.6 85.1  0.7 ND ND ND ND ND Bleomycin-bound DSC 94.7  0.8 100.3  0.8 ND ND ND ND ND Wild type. Circular dichroism spectroscopy ( ). 205 nm Fluorescence spectroscopy ( ,  ). ex 295 nm em 315 nm ND, not determined. TABLE III Data collection and refinement statistics Values in parentheses refer to the highest resolution shell. X-ray data collection statistics Wavelength (Å) 0.90830 Resolution (Å) 30–1.5 (1.53–1.50) Total observations 87,740 (3,602) Unique observations 36,444 (1,714) Completeness (%) 94.4 (88.6) I/(I) 20.3 (3.9) R 0.043 (0.189) sym Refinement statistics Resolution (Å) 30.0–1.5 R 0.174 FIG.2. DNA protection assay. Digital photographs of 1% agarose work R 0.194 free gels showing the degree of DNA protection by Shble variants against r.m.s.d. bond distances (Å) 0.01 the strand scission action of bleomycin A2 are shown. A, assay at 25 °C r.m.s.d. bond angles (°) 1.664 for 10 min. B, assay at 25 °C for 10 min after protein preincubation at Total number of non-hydrogen atoms 1,883 85 °C for 30 min. C, assay at 85 °C for 10 min. ble, bleomycin A2; WT, Average protein B-value (Å ) 13.3 wild type. Number of solvent molecules 288 Average solvent B-value (Å ) 27.0 3C. The structure forms a compact, homodimeric / protein of Number of bleomycin atoms 182 Average ligand B-value (Å ) 22.8 121 amino acids (Met-1, Gln-123, and Asp-124 are disordered) Number of sulfate ion atoms 15 in which two bleomycin A2 molecules are accommodated in Average sulfate ion B-value (Å ) 31.4 binding pockets at the dimer interface. These pockets consist of R  I  I / I where I is the intensity of a given sym hkl i i hkl i i i a hydrophilic concavity that runs into a hydrophobic intersub- measurement, and the sums are over all measurements and reflections. unit crevice. The dimer is maintained by alternate N-terminal b R  F   F / F  for 95% of the reflection data used in work obs calc obs -strand hydrogen bonding between both monomers and by refinement. R  F   F / F  for the remaining 5%. Van der Waals interactions at the largely hydrophobic subunit free obs calc obs contact (19, 45). Three sulfate ions are present at the surface of the dimer of which two form ion pairs with Arg-104 of both (r.m.s.d., 0.38 Å; Table IV) only revealed large differences in a chains. The presence of a dimer in the asymmetric unit allowed random coil region comprising residues Asp-88, Ala-89, and the identification of certain symmetry deviations between both Ser-90 (Fig. 3, A and B) that is spatially close to the carbamoyl monomers. A backbone superimposition of both chains group of the D-mannose moiety of bleomycin (Fig. 4A). Their Crystal Structure of Thermostable Bleomycin-binding Protein 11427 FIG.3. Structure and electron density of HTS in complex with bleomycin A2. A ribbon diagram showing the dimeric structure of the 4-fold mutant Shble in complex with bleomycin A2 is shown. Mutations are indicated by stick representations. Chain A is in blue, chain B is in red, G18E is in green, D32V is in pink, L63Q is in yellow, and G98V is in orange. A, side view. B, viewed from the N- and C-terminal side (top view). C, stereoview of the electron density around residues Pro-9 and Trp-65 contoured at 2 . Residues are colored according to the Corey-Pauling- Koltun color scheme, and water molecules are represented by red spheres. TABLE IV Crystal structures of bleomycin-binding proteins Protein Data Protein/source Form Resolution Backbone r.m.s.d. Ref. Bank code ÅÅ 1BYL Uncomplexed 2.30 0.63 19 Shble S. hindustanus 1XRK Bleomycin A2-bound 1.50 0.38 This study BlmA S. verticillus 1QTO Uncomplexed 1.50 0.77 45 1JIE Bleomycin A2-bound 1.80 0.73 23 2 a 1JIF Cu -bleomycin A2-bound 1.60 0.75 23 1ECS Uncomplexed 1.70 1.18 22 BlmT K. pneumoniae transposon Tn5 1EWJ Bleomycin A2-bound 2.50 1.15 22 Averaged backbone superimposition r.m.s.d. values compared with HTS chains A and B. Backbone superimposition r.m.s.d. values of HTS chain A to B. respective C atoms deviate 2.0, 5.1, and 1.6 Å in position while standing of these proteins in general. Moreover the structure giving rise to almost oppositely pointing amino acid side has revealed several molecular features that can account for chains. In contrast to the bleomycin-bound and unbound BlmA increased protein stability and improved functionality at structure, backbone B-factors in this region are only margin- higher temperature in vivo and in vitro. ally higher compared with the average value, suggesting a rigid Structural Effects of Mutations conformation (23, 45). The difference in orientation of this loop might therefore be the result of sequential binding of two Introduction of an Intersubunit Hydrogen Bond Network— bleomycin molecules. Unlike BlmA, no symmetry-related dif- The structure of the dimer shows that each of the two bleomycin ferences were observed in the region between amino acids 100 A2 molecules is bound by the concerted action of 21 amino acids. and 103. The topology of the HTS protein complex and the Due to its intersubunit location, both binding sites are composed mode of bleomycin binding are similar to other BBPs. An over- of residues from either subunit. These include Val-32, Phe-33, view of available structures is given in Table IV. Glu-35, Phe-38, Ser-51, Ala-52, and Val-53 of one subunit and The structure of the HTS mutant in complex with bleomycin Pro-59, Asp-60, Asn-61, Thr-62, Gln-63, Trp-65, Phe-86, Ala-89, completes the list of structural information of three BBPs with Trp-102, Ala-107, Arg-109, Gly-113, Cys-115, and His-117 of the and without their ligands, hereby contributing to our under- other. The crystal structure clearly reveals the central role of 11428 Crystal Structure of Thermostable Bleomycin-binding Protein FIG.4. Chemical diagram and electron density of bleomycin A2. A, schematic representation of bleomycin A2. B, electron density around bleomycin A2 contoured at 1.5 . The diagram indicates the missing electron density around the -aminopropyldimethylsulfonium moiety suggesting a disordered conformation. mutation L63Q, which was found in three of five different double ing hydrophobic packing of surface indentations, thereby rein- mutants. Gln-63 is involved in an extensive hydrogen bond net- forcing subsurface secondary structure elements (48). This work at the bottom of the bleomycin binding concavity (Fig. 5A). might explain the enhanced secondary structure preservation It is noteworthy that the carbonyl side chains (O-1) of both at high temperatures as inferred from CD of mutant 77-3. Gln-63 residues in the dimer act as terminal hydrogen bond Strikingly modeling of G40A into the wild-type structure (not acceptors of a five-molecule water channel present at the dimer shown) revealed close spatial proximity to Asp-32 (5 Å between interface. A second hydrogen bond is accepted from the side chain C and 3.5 Å between C atoms) that may also underline a hydroxyl group (O-) of Ser-51 of the adjacent subunit. The amide similar need for hydrophobicity in this part of the protein. side chain (N-2) of Gln-63 forms a hydrogen bond with one of two Mutation Arg-31 to Leu, which occurred in two types of double water molecules trapped between the bleomycin and the surface mutants, came as a surprise since it is involved in a surface ion of the protein. The presence of a leucine at position 63 would most pair with Asp-25 in wild-type Shble. Apparently this electro- likely not have allowed for a hydrogen bond network of this size. static interaction does not counterweight the beneficial effects The advantage of an amino acid compatible with hydrogen bond- of improved hydrophobic packing among residues Val-20, Thr- ing at position 63 is also evident from the alignment, which 24, Val-34, and Val-41. indicates that without exception the other four BBPs have a Reduction of Surface Loop Flexibility—From previous crys- serine at this specific site (Fig. 1). Although the mutant structure tallographic and NMR studies of BBPs, it has become clear that without bleomycin is not available, we speculate that an inter- the loop following Gly-98 in Shble will change its conformation subunit hydrogen bond between Gln-63 and Ser-51 can persist upon binding of the antibiotic (22, 23, 45, 46). This conforma- even without the antibiotic bound, giving rise to a beneficial tional change enables the tryptophan at position 102 to stack interaction that might stabilize the dimer at high temperatures. optimally with the hydrophobic bithiazole moiety of bleomycin Hydrophobic Packing of Surface Indentations—In wild-type (Fig. 4A), packing both thiazole rings tightly against Phe-33 Shble and BlmA, Asp-32 is located on the edge of the intersub- and Phe-38 of the adjacent subunit. In both BlmA and Shble, unit binding groove for the bithiazole moiety and tail region of Gly-98, located on the edge of a small -strand leading toward bleomycin (Fig. 4A) (19, 23, 45). In bleomycin A2 and B2 and in the binding loop, seems to have a hinge function (Fig. 5C). The phleomycin D1 the tail is positively charged, suggesting in- bending motion of the backbone is also clearly reflected in large volvement of Asp-32 in electrostatic stabilization or ligand  and torsion angle changes of more than 20° upon the recognition. From NMR studies it has become clear that in the binding of bleomycin. At high temperatures, however, this flex- bound state no strong interactions occur between the protein ibility might have caused problems leading to local unfolding or and the positively charged tail of bleomycin (46). These data a decreased bleomycin binding ability. A substitution for either are supported by the absence of electron density for the -amin- a valine or serine as observed would increase the rigidity of the opropyldimethylsulfonium moiety of bleomycin A2 in the bi- loop and could therefore restore the binding capacity at nary complex structure of BlmA and HTS, suggesting a disor- high temperature. dered conformation of the tail end (Fig. 4B) (23). This might -Helix Stabilization—Mutation G18E introduces a gluta- have allowed for an amino acid substitution to valine, which mate at position N-3 in the first turn of the largest -helix of the extends the hydrophobic bithiazole binding cleft at the dimer protein. Statistical analysis and experimental studies have interface fitting nicely within a highly hydrophobic environ- shown that glutamates are energetically highly favored over ment consisting of Phe-33, Phe-38, Val-42, Thr-47, and Phe-49 glycines at the third position in an -helix (49, 50). This effect is (Fig. 5B). In addition, both BlmT and BlmS sequences also most likely caused by the stabilizing effect of the negatively contain a valine at the corresponding position (Fig. 1). From a charged side chain on the helix macrodipole. To our surprise, thermodynamic point of view, a mutation introducing surface chain A of the crystal structure revealed the formation of a hydrophobicity is generally believed to be unfavorable and has genuine i, i  5 -helix surface ion pair between Glu-21 and therefore rarely been investigated in directed mutagenesis Arg-26 that was absent in the wild-type structure (Fig. 5D). This studies. Nevertheless some studies have reported significant new ion pair may have been the result of repulsion of anionic improvements in protein stability by placing bulky hydropho- glutamate side chains of positions 18 and 21, directing the latter bic amino acids at the surface of a neutral protease from toward the C-terminal arginine. Although i, i  5 -helical sur- G. stearothermophilus (47). Recent findings using Bacillus li- face ion pairs do not give rise to strong ionic interactions at cheniformis -amylase have clearly indicated that hydrophobic ambient temperatures (51), they might be more favorable at surface residues can indeed be extremely stabilizing by improv- higher temperatures. Theoretical models have indicated that the Crystal Structure of Thermostable Bleomycin-binding Protein 11429 FIG.5. Structural effects of the individual mutations. A, L63Q. A ribbon diagram showing the hydrogen bond network at the dimer interface is shown. Thr-62, Gln-63, Ser-51, and bleomycin A2 are shown together with the intersubunit water channel. B, D32V. A ribbon diagram showing the hydrophobic intersubunit bleomycin tail binding crevice is shown. Val-32 may be involved in improved hydrophobic packing of this surface indentation among amino acids Phe-33, Phe-38, Val-42, Thr-47, and Phe-49. C, G98V. A ribbon representation showing a loop between Pro-92 and Pro-111 that is involved in bleomycin binding is shown. Val-98 is located at a former hinge region that enables Trp-102 to stack the bithiazole tail against Phe-33 and Phe-38. The electron density revealed two alternative side chain rotamers for Val-98 in chain A (not shown) and a single side chain conformation in chain B. D, G18E. A side-by-side comparison of -helix 1 formed between Asp-15 and Leu-27 in the wild-type and mutant crystal structures of Shble. The existence of a surface ion pair between Glu-21 and Arg-26 is visible in the electron density (contoured at 1.5 ). Interatomic distances are indicated in Å. Corey-Pauling-Koltun color coding was used for amino acids and bleomycin A2. Chain A is indicated in blue, and chain B is indicated in red. Water molecules are represented by red spheres. Hydrogen bonds are depicted by green dotted lines. energetic cost of desolvating charged groups is much less at residues (Asn, Gln, Ser, and Thr) (54). This is fully in agree- 100 °C due to a drop in the dielectric constant of water (52). This ment with the requirements for the elevated numbers of sur- is currently the best explanation for the fact that proteins from face salt bridges and improved hydrophobic core packing that extreme thermophiles have large ion pair networks at their sur- has generally been recognized in these types of proteins. These faces that are thought to be involved in maintaining structural are just two of a multitude of mechanisms that proteins from integrity (3). Additionally a minor beneficial effect of this muta- hyperthermophilic microorganisms have used to deal with ex- tion could be the introduction of additional negative surface treme temperatures (2, 55, 56). charge, which enhances electrostatic attraction of the cationic Recently several other random mutagenesis studies have antibiotic under physiological conditions. also reported large improvements in thermostability by apply- Laboratory Versus Natural Evolution of Thermostability—In ing directed evolution approaches. A mesophilic xylanase of this study, several possible mechanisms of adaptation to high family 11 was stabilized by over 35 °C by combining nine mu- temperature were identified, such as the introduction of a tations found separately after extensive screening. The activity hydrogen bond network, improved hydrophobic packing of sur- of this mutant was optimized by saturation mutagenesis of all face indentations, reduction of loop flexibility, and -helix sta- mutated positions, yielding an enzyme variant with highly bilization. Remarkably half of all mutations found were glycine enhanced properties for high temperature applications (8). In replacements, which could point to protein stabilization by another study, a highly thermostable esterase containing seven decreasing the entropy of the unfolded state (53). Although this mutations was evolved in six rounds of random mutagenesis, could be a general strategy of stabilization, proteins from hy- recombination, and screening (7). The resulting enzyme was perthermophiles do not have a lower glycine content than their crystallized, and its structure was determined (57). The struc- mesophilic counterparts but rather a slightly increased one ture revealed that improved stability was due to altered core (54). Their predicted proteomes do have an increased propen- packing, -helix stabilization, the introduction of surface salt sity for charged (Arg, Lys, and Glu) and bulky aliphatic (Ile and bridges, and reduction of flexibility in surface loops. From these Val) amino acids that has mostly come at the cost of polar and many other directed evolution and site-directed mutagen- 11430 Crystal Structure of Thermostable Bleomycin-binding Protein Acknowledgments—We thank Dr. J. Berenguer for helpful sugges- esis studies, it has become apparent that (i) proteins can be tions and Anton Korteweg for technical assistance with DSC. stabilized substantially by small numbers of mutations, (ii) these mutations are often located at the protein surface, and REFERENCES (iii) their effects are usually additive. As few as two of 12 amino 1. Arnold, F. H., Wintrode, P. L., Miyazaki, K., and Gershenson, A. (2001) Trends acid differences between a mesophilic and thermophilic cold Biochem. Sci. 26, 100–106 2. Vieille, C., and Zeikus, G. J. (2001) Microbiol. Mol. Biol. Rev. 65, 1–43 shock protein turned out to be responsible for the difference in 3. Karshikoff, A., and Ladenstein, R. (2001) Trends Biochem. Sci. 26, 550–556 thermostability (58). The remaining variation in sequence 4. Serrano, L., and Fersht, A. R. (1989) Nature 342, 296–299 might just have occurred as a result of neutral sequence drift or 5. Van den Burg, B., Vriend, G., Veltman, O. R., Venema, G., and Eijsink, V. G. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 2056–2060 specific properties required by the host, such as solubility, 6. Stemmer, W. P. (1994) Nature 370, 389–391 turnover, and molecular interactions (59). 7. Giver, L., Gershenson, A., Freskgard, P. O., and Arnold, F. H. (1998) Proc. Despite the fact that only a small number of mutations are Natl. Acad. Sci. U. S. A. 95, 12809–12813 8. Palackal, N., Brennan, Y., Callen, W. N., Dupree, P., Frey, G., Goubet, F., required to render a protein thermostable, finding those muta- Hazlewood, G. P., Healey, S., Kang, Y. E., Kretz, K. A., Lee, E., Tan, X., tions remains a difficult task. Apart from screening vast num- Tomlinson, G. L., Verruto, J., Wong, V. W., Mathur, E. J., Short, J. M., Robertson, D. E., and Steer, B. A. (2004) Protein Sci. 13, 494–503 bers of random mutants in microtiter plates, in vitro and in 9. Liao, H., McKenzie, T., and Hageman, R. (1986) Proc. Natl. Acad. Sci. U. S. A. vivo selection schemes offer great advantages to reduce the 83, 576–580 10. Tamakoshi, M., Yamagishi, A., and Oshima, T. (1995) Mol. Microbiol. 16, effort to encounter improved variants. Currently two main 1031–1036 strategies are available. 11. Hoseki, J., Yano, T., Koyama, Y., Kuramitsu, S., and Kagamiyama, H. (1999) One of these thermostabilizing selection methods is called J. Biochem. (Tokyo) 126, 951–956 12. Friedrich, A., Prust, C., Hartsch, T., Henne, A., and Averhoff, B. (2002) Appl. Proside (protein stability increased by directed evolution) Environ. Microbiol. 68, 745–755 which is based on the empirically derived inverse correlation 13. Noll, K. M., and Vargas, M. (1997) Arch. Microbiol. 168, 73–80 between protein thermostability and proteolytic susceptibility 14. Gatignol, A., Durand, H., and Tiraby, G. (1988) FEBS Lett. 230, 171–175 15. Sugiyama, M., and Kumagai, T. (2002) J. Biosci. Bioeng. 93, 105–116 (60). In a phage display-like procedure, a library of mutants is 16. Poddevin, B., Orlowski, S., Belehradek, J., Jr., and Mir, L. M. (1991) Biochem. fused between two domains of E. coli filamentous phage Fd Pharmacol. 42, (suppl.) S67–S75 17. Sikic, B. I. (1986) Cancer Surv. 5, 81–91 gene-3-protein, which is then subjected to proteases while in- 18. Sugiyama, M., Thompson, C. J., Kumagai, T., Suzuki, K., Deblaere, R., Villar- ducing local unfolding of the target protein by means of tem- roel, R., and Davies, J. (1994) Gene (Amst.) 151, 11–16 perature or chemical denaturants. The resulting unstable fu- 19. Dumas, P., Bergdoll, M., Cagnon, C., and Masson, J. M. (1994) EMBO J. 13, 2483–2492 sion protein variants are cleaved, causing only the surviving, 20. Genilloud, O., Garrido, M. C., and Moreno, F. (1984) Gene (Amst.) 32, 225–233 more stable phages to be found after infection. It was found 21. Semon, D., Movva, N. R., Smith, T. F., el Alama, M., and Davies, J. (1987) Plasmid 17, 46–53 that in an ionic denaturant, non-polar surface interactions 22. Maruyama, M., Kumagai, T., Matoba, Y., Hayashida, M., Fujii, T., Hata, Y., were optimized, whereas at elevated temperature variants and Sugiyama, M. (2001) J. Biol. Chem. 276, 9992–9999 with improved surface electrostatics were selected (59). 23. Sugiyama, M., Kumagai, T., Hayashida, M., Maruyama, M., and Matoba, Y. (2002) J. Biol. Chem. 277, 2311–2320 Cloning and selection in a thermophile, as conducted in this 24. Drocourt, D., Calmels, T., Reynes, J. P., Baron, M., and Tiraby, G. (1990) study, is a second directed evolution strategy that can lead to Nucleic Acids Res. 18, 4009 25. Nuttall, S. D., Deutschel, S. E., Irving, R. A., Serrano-Gomicia, J. A., and rapid improvements in thermostability. Despite its simplicity, Dyall-Smith, M. L. (2000) Biochem. J. 346, 251–254 very few studies using this technique have been reported. This 26. Boyer, H. W., and Rouland-Dussoix, R. (1969) J. Mol. Biol. 41, 459–472 is most likely due to its limited applicability because the gene 27. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd Ed., Cold Spring Harbor Laboratory, Cold Spring of interest rarely confers any biologically relevant function for Harbor, NY growth or survival of the thermophile. Nonetheless function- 28. de Grado, M., Castan, P., and Berenguer, J. (1999) Plasmid 42, 241–245 29. Shafikhani, S., Siegel, R. A., Ferrari, E., and Schellenberger, V. (1997) Bio- ally stabilized mutants have been reported up to 79 °C for a Techniques 23, 304–310 kanamycin nucleotidyltransferase (9, 11) and 58 °C for a chlor- 30. Koyama, Y., Hoshino, T., Tomizuka, N., and Furukawa, K. (1986) J. Bacteriol. amphenicol acetyltransferase (61). Auxotrophic knock-out 166, 338–340 31. Stemmer, W. P., Crameri, A., Ha, K. D., Brennan, T. M., and Heyneker, H. L. strains for leucine biosynthesis were complemented with ther- (1995) Gene (Amst.) 164, 49–53 molabile counterpart genes from B. subtilis and Saccharomyces 32. Bradford, M. M. (1976) Anal. Biochem. 72, 248–254 33. Rondeau, J. M., Cagnon, C., Moras, D., and Masson, J. M. (1989) J. Mol. Biol. cerevisiae and adapted to higher temperature by serial accu- 207, 645–646 mulation of beneficial mutations in the in trans introduced 34. Otwinowski, A., and Minor, W. (1997) Methods Enzymol. 276, 307–326 3-isopropylmalate dehydrogenase genes (62, 63). A hybrid -ga- 35. Vagin, A. A., and Teplyakov, A. (1997) J. Appl. Crystallogr. 30, 1022–1025 36. Perrakis, A., Morris, R., and Lamzin, V. S. (1999) Nat. Struct. Biol. 6, 458–463 lactosidase consisting of Bacillus stearothermophilus and T. 37. Murshudov, G. N., Vagin, A. A., and Dodson, E. J. (1997) Acta Crystallogr. thermophilus peptide regions was adapted to function at 67 °C Sect. D. Biol. Crystallogr. 53, 240–255 by selection for growth on melibiose as the sole carbon and 38. Laskowski, R. A., MacArthur, M. W., Moss, D. S., and Thornton, J. M. (1993) J. Appl. Crystallogr. 26, 283–291 energy source (64). 39. Vaguine, A. A., Richelle, J., and Wodak, S. J. (1999) Acta Crystallogr. Sect. D Biol. Crystallogr. 55, 191–205 40. Guex, N., and Peitsch, M. C. (1997) Electrophoresis 18, 2714–2723 Concluding Remarks 41. Fernandez-Herrero, L. A., Olabarria, G., and Berenguer, J. (1997) Mol. Micro- biol. 24, 61–72 In this study, several double mutants of a bleomycin-binding 42. Tomita, K., Nakakita, Y., Hoshino, Y., Numata, K., and Kawaguchi, H. (1987) Int. J. Syst. Bacteriol. 37, 211–213 protein were isolated with enhanced performance at high tem- 43. Jaswal, S. S., Sohl, J. L., Davis, J. H., and Agard, D. A. (2002) Nature 415, perature both in vivo and in vitro. Structural analysis showed 343–346 that the mutations gave rise to different means of stabilization 44. Gonzalez, M., Argarana, C. E., and Fidelio, G. D. (1999) Biomol. Eng. 16, 67–72 45. Kawano, Y., Kumagai, T., Muta, K., Matoba, Y., Davies, J., and Sugiyama, M. in four parts of the protein. A combined mutant gene with a low (2000) J. Mol. Biol. 295, 915–925 GC content was created that can serve as an antibiotic resist- 46. Vanbelle, C., Brutscher, B., Blackledge, M., Muhle-Goll, C., Remy, M. H., Masson, J. M., and Marion, D. (2003) Biochemistry 42, 651–663 ance marker for aerobic and microaerophilic mesophiles, ther- 47. Van den Burg, B., Dijkstra, B. W., Vriend, G., Van der Vinne, B., Venema, G., mophiles, and hyperthermophiles. This may allow the develop- and Eijsink, V. G. (1994) Eur. J. Biochem. 220, 981–985 48. Machius, M., Declerck, N., Huber, R., and Wiegand, G. (2003) J. Biol. Chem. ment of efficient shuttle vectors and knock-out strategies for 278, 11546–11553 hyperthermophilic Archaea and bacteria based on positive se- 49. Iqbalsyah, T. M., and Doig, A. J. (2004) Protein Sci. 13, 32–39 lection schemes. Moreover the high GC content double mutant 50. Penel, S., Hughes, E., and Doig, A. J. (1999) J. Mol. Biol. 287, 127–143 51. Smith, J. S., and Scholtz, J. M. (1998) Biochemistry 37, 33–40 genes will now permit multigene knock-out strategies in ther- 52. Elcock, A. H. (1998) J. Mol. Biol. 284, 489–502 mophiles such as T. thermophilus, allowing further exploration 53. Van den Burg, B., and Eijsink, V. G. (2002) Curr. Opin. Biotechnol. 13, and exploitation of thermophilic microbial sources. 333–337 Crystal Structure of Thermostable Bleomycin-binding Protein 11431 54. Cambillau, C., and Claverie, J. M. (2000) J. Biol. Chem. 275, 32383–32386 61. Turner, S. L., Ford, G. C., Mountain, A., and Moir, A. (1992) Protein Eng. 5, 55. Kumar, S., and Nussinov, R. (2001) Cell Mol. Life Sci. 58, 1216–1233 535–541 56. Vogt, G., Woell, S., and Argos, P. (1997) J. Mol. Biol. 269, 631–643 62. Akanuma, S., Yamagishi, A., Tanaka, N., and Oshima, T. (1998) Protein Sci. 7, 57. Spiller, B., Gershenson, A., Arnold, F. H., and Stevens, R. C. (1999) Proc. Natl. 698–705 Acad. Sci. U. S. A. 96, 12305–12310 63. Tamakoshi, M., Nakano, Y., Kakizawa, S., Yamagishi, A., and Oshima, T. 58. Perl, D., Mueller, U., Heinemann, U., and Schmid, F. X. (2000) Nat. Struct. (2001) Extremophiles 5, 17–22 Biol. 7, 380–383 64. Fridjonsson, O., Watzlawick, H., and Mattes, R. (2002) J. Bacteriol. 184, 59. Martin, A., Sieber, V., and Schmid, F. X. (2001) J. Mol. Biol. 309, 717–726 3385–3391 60. Sieber, V., Pluckthun, A., and Schmid, F. X. (1998) Nat. Biotechnol. 16, 65. Thompson, J. D., Gibson, T. J., Plewniak, F., Jeanmougin, F., and Higgins, 955–960 D. G. (1997) Nucleic Acids Res. 25, 4876–4882 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Biological Chemistry American Society for Biochemistry and Molecular Biology

Engineering a Selectable Marker for Hyperthermophiles *

Download PDF
10 pages

Loading next page...
 
/lp/american-society-for-biochemistry-and-molecular-biology/engineering-a-selectable-marker-for-hyperthermophiles-wFSbrF69h0

References (65)

Publisher
American Society for Biochemistry and Molecular Biology
Copyright
Copyright © 2005 Elsevier Inc.
ISSN
0021-9258
eISSN
1083-351X
DOI
10.1074/jbc.m413623200
Publisher site
See Article on Publisher Site

Abstract

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 280, No. 12, Issue of March 25, pp. 11422–11431, 2005 © 2005 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. □ S Received for publication, December 3, 2004, and in revised form, December 27, 2004 Published, JBC Papers in Press, January 7, 2005, DOI 10.1074/jbc.M413623200 Stan J. J. Brouns‡§, Hao Wu‡, Jasper Akerboom‡, Andrew P. Turnbull , Willem M. de Vos‡, and John van der Oost‡ From the ‡Laboratory of Microbiology, Department of Agrotechnology and Food Sciences, Wageningen University, Hesselink van Suchtelenweg 4, 6703 CT Wageningen, The Netherlands and Proteinstrukturfabrik c/o Berliner Elecktronenspeicherring-Gesellschaft fu¨r Synchrotronstrahlung GmbH, Albert Einstein Strasse 15, D-12489 Berlin, Germany above 80 °C, improving the thermal stability of a protein is still Limited thermostability of antibiotic resistance mark- ers has restricted genetic research in the field of ex- a challenging task. This is mainly because the laws governing tremely thermophilic Archaea and bacteria. In this protein stability are not easily extracted because they are study, we used directed evolution and selection in the highly variable and complex (1, 2). It seems generally accepted thermophilic bacterium Thermus thermophilus HB27 to that the extreme stability of certain natural proteins results find thermostable variants of a bleomycin-binding pro- from the cumulative effect of small adaptations in protein tein from the mesophilic bacterium Streptoalloteichus architecture and amino acid composition. Although some of hindustanus. In a single selection round, we identified these stabilizing features, such as optimized surface ion pair eight clones bearing five types of double mutated genes networks (3), are unlikely to be engineered into a protein of that provided T. thermophilus transformants with bleo- interest, other strategies like -helix capping (4) and the intro- mycin resistance at 77 °C, while the wild-type gene could duction of disulfide bonds and prolines in -turns (5) can be only do so up to 65 °C. Only six different amino acid applied very successfully when carefully designed on the basis positions were altered, three of which were glycine res- of a high resolution crystal structure. However, in many cases idues. All variant proteins were produced in Esche- atomic resolution three-dimensional information of a protein is richia coli and analyzed biochemically for thermal sta- unavailable. Directed evolution approaches, by contrast, do not bility and functionality at high temperature. A synthetic require any structural information and commonly rely on ran- mutant resistance gene with low GC content was de- signed that combined four substitutions. The encoded dom mutagenesis and recombination followed by screening or protein showed up to 17 °C increased thermostability selection schemes (1, 6). Thermostability screens of mutant and unfolded at 85 °C in the absence of bleomycin, libraries are usually carried out by applying a thermal chal- whereas in its presence the protein unfolded at 100 °C. lenge at nonpermissive temperatures after which the remain- Despite these highly thermophilic properties, this mu- ing functionality of the individual clones is tested (7, 8). To tant was still able to function normally at mesophilic explore sufficient sequence space requires the testing of large temperatures in vivo. The mutant protein was co-crys- numbers of mutant clones, which necessitates high throughput tallized with bleomycin, and the structure of the binary approaches such as the use of robotics. Conversely efficient complex was determined to a resolution of 1.5 Å. De- selection procedures allow the testing of a large set of variants tailed structural analysis revealed possible molecular while reducing the effort of finding improved ones to mechanisms of thermostabilization and enhanced anti- a minimum. biotic binding, which included the introduction of an A convenient selection system for finding protein variants in intersubunit hydrogen bond network, improved hydro- a library with improved thermostability is based on in vivo phobic packing of surface indentations, reduction of screening in a thermophilic expression host. Cloning and selec- loop flexibility, and -helix stabilization. The potential tion in thermophilic microorganisms such as Geobacillus applicability of the thermostable selection marker is discussed. stearothermophilus (30–60 °C) or Thermus thermophilus (50– 80 °C) mimic natural evolution but are only applicable when the gene of interest encodes a protein that is of biological Despite the vast amount of protein sequences and structures relevance to growth or survival of the host organism (9, 10). from microorganisms that grow optimally at temperatures The selective pressure can be fine tuned by raising the temper- ature of growth, enabling only hosts that bear thermoadapted * This work was supported by a grant from the European Union in variants to grow on solid media. For instance, a combination of the framework of the SCREEN project (Contract QLK3-CT-2000- in vitro mutagenesis methods and in vivo selection schemes 00649). The costs of publication of this article were defrayed in part by have led to a highly thermostable kanamycin nucleotidyltrans- the payment of page charges. This article must therefore be hereby ferase gene that is able to function at temperatures up to 79 °C marked “advertisement” in accordance with 18 U.S.C. Section 1734 (11). Such mutant selection markers have permitted the devel- solely to indicate this fact. □ S The on-line version of this article (available at http://www.jbc.org) opment of genetic tools that are very useful in the study of contains Supplemental Table 1. gene-function relationships in thermophilic bacteria (12). The nucleotide sequence(s) reported in this paper has been submitted TM In contrast to thermophiles (optimum temperature for to the GenBank /EBI Data Bank with accession number(s) AY780486. The atomic coordinates and structure factors (code 1XRK) have been growth, 60–80 °C), antibiotic-based genetic systems for hyper- deposited in the Protein Data Bank, Research Collaboratory for Struc- thermophilic bacteria and Archaea (optimum temperature for tural Bioinformatics, Rutgers University, New Brunswick, NJ growth, 80 °C) are still in their infancy. This is primarily due (http://www.rcsb.org/). to the absence of thermostable antibiotics and their corre- § To whom correspondence should be addressed. Tel.: 31-317-483110; Fax: 31-317-483829; E-mail: [email protected]. sponding resistance factors because most known antibiotic- 11422 This paper is available on line at http://www.jbc.org This is an Open Access article under the CC BY license. Crystal Structure of Thermostable Bleomycin-binding Protein 11423 kanamycin nucleotidyltransferase gene. Ligation mixtures were trans- producing microorganisms are mesophilic bacteria and fungi. formed into E. coli HB101, and transformants were plated on 1.5% LB Often the common antibiotics cannot be used because many of agar plates supplemented with 3 g/ml bleomycin. Both blmS and shble them are unstable at high temperatures, or hyperthermophiles provided resistance against the antibiotic, giving rise to the 4434-bp are simply insensitive to them (13). The glycopeptide bleomycin plasmid pWUR111 and the 4404-bp plasmid pWUR112, respectively. is an exception since it is a highly thermostable molecule and Mutant Library Construction—Error-prone PCR was carried out us- effective against many aerobic microorganisms and eukaryotic ing two different polymerases, namely Taq (Amersham Biosciences) and Mutazyme (Genemorph kit, Stratagene). This approach was chosen cell lines (14, 15). The bleomycin family of antibiotics, including to complement the transition and transversion bias of each enzyme to phleomycin and tallysomycin, are DNA- and RNA-cleaving gly- provide a more complete mutational spectrum of the PCR product. For copeptides that are produced by the actinomycetes Strep- the error-prone amplification, flanking primers BG1412 (sense, 5- toalloteichus hindustanus and Streptomyces verticillus. As lit- CGACCCTTAAGGAGGTGTGAGGCATATG-3) and BG1408 (anti- tle as a few hundred bleomycin molecules can effectively kill sense, 5-CGAGCTCGGTACCCGGGGATCCTCTAGATTA-3) (NdeI aerobic cells (16). For this reason, bleomycin is currently clin- and XbaI sites are underlined) were designed to allow variation throughout the entire coding sequence between the start and stop codon ically used as an antitumor agent against squamous cell carci- (indicated in boldface). Taq polymerase-based PCRs were performed as nomas and malignant lymphomas (17). Resistance against described previously (29). A 50-l PCR contained 5 ng of pWUR112, 5 bleomycin-like antibiotics is conferred by N-acetylation, deami- pmol of each primer, 0.2 mM dATP and dGTP, 1 mM dCTP and dTTP, 5 dation, and sequestration of the molecule (15). The latter mech- units of polymerase, 3 mM MgCl , and three concentrations of MnCl 2 2 anism involves bleomycin-binding proteins (BBPs), which (0.1, 0.3, and 0.5 mM). The mixture was thermocycled as follows: 95 °C have been found only in mesophilic bacteria. Two proteins, (4 min); 30 cycles of 94 °C (30 s), 55 °C (45 s), and 72 °C (25 s); and postdwelled for 4 min at 72 °C. Mutazyme PCRs were prepared accord- Shble and BlmA, provide self-immunity for bleomycin produc- ing to the manufacturer’s instructions and thermocycled as above using ers S. hindustanus and S. verticillus, respectively (14, 18), and an elongation time of 50 s. may be involved the transport and excretion of the molecule Randomly mutated PCR products were cloned into vector pMK18 (19). Two genes, blmT and blmS, are located on the Klebsiella and transformed into E. coli HB101. A total of 10,000 Taq- and 10,000 pneumoniae transposon Tn5 (20) and on the Staphylococcus Mutazyme-derived clones were resuspended in 50 ml of LB medium aureus plasmid pUB110 (21), respectively. All four proteins are supplemented with antibiotic and grown in 1 liter of medium to early stationary phase. Plasmids were subsequently harvested using a Mini- highly negatively charged cytoplasmic proteins of around 14 prep plasmid isolation kit (Qiagen). kDa that form homodimers that bind two positively charged Selection in T. thermophilus—T. thermophilus HB27 was kindly antibiotic molecules at a hydrophobic subunit interface cleft provided by Dr. J. Berenguer (Autonomous University of Madrid, Ma- (19, 22, 23). The small protein size and the wide applicability of 2 drid, Spain). Cells were routinely cultivated at 70 °C in a Ca - (3.9 mM) the drug have made both shble and blmT popular dominant and Mg (1.9 mM)-rich medium (28) containing 8 g/liter tryptone, 4 selection markers in vector systems for lower and higher eu- g/liter yeast extract, and 3 g/liter NaCl dissolved in Evian mineral water (pH 7.7, after autoclaving) (Evian-les-Bains, France). Transfor- karyotes, bacteria, and halophilic Archaea (15, 24, 25). This mation of T. thermophilus was essentially performed by the method of prompted us to investigate whether we could thermostabilize Koyama (30). Frozen cell aliquots were resuspended in 25 ml of medium Shble and BlmS to allow for their application in aerobic ther- and grown at 150 rpm to an A of 0.8. The culture was then diluted 1:1 mophiles and hyperthermophiles. in preheated medium and incubated for another hour. Next plasmids In this study, we performed directed evolution using selec- were added to 0.5 ml of culture, and the mixture was incubated for 2–3 tion in the thermophilic bacterium T. thermophilus and ob- h at 70 °C with occasional shaking before being plated on 3% agar plates (BD Biosciences) supplemented with 30 g/ml kanamycin or 15 tained various mutant proteins that could operate under highly g/ml bleomycin (Calbiochem) for selection. Colonies appeared within thermophilic growth conditions. Their enhanced performance 36 h at 60–70 °C. At temperatures above 70 °C, 1% Gelrite plates (Roth, at high temperature was analyzed biochemically, and possible Karlsruhe, Germany) were used supplemented with 100 g/ml kana- stabilizing effects were identified. mycin or 20 g/ml bleomycin for selection. Colonies were grown over- night in liquid medium containing 30 g/ml kanamycin or 5 g/ml EXPERIMENTAL PROCEDURES bleomycin. T. thermophilus plasmid DNA was prepared using a plasmid All chemicals were of analytical grade and purchased from Sigma. Miniprep kit (Qiagen) after a preincubation with 2 mg/ml lysozyme for Primers were obtained from MWG Biotech AG (Ebersberg, Germany). 30 min at 37 °C. Polymerase chain reactions were performed with Pfu TURBO (Strat- Gene Cloning, Overexpression, and Protein Purification—Wild-type agene) unless stated otherwise. Bleomycin A2 (Bleocin, Calbiochem) and double mutant shble genes were PCR-amplified from their respec- was used for all selections. Escherichia coli HB101 (F hsdS20 (r , tive pWUR112 plasmids using primers BG1503 (sense, 5-GATGGC- m ) ara-14 galK2 lacY1 leuB6 mcrB mtl-1 proA2 recA13 rpsL20 CATGGCCAAGTTGACCAGTGC-3) and BG1504 (antisense, 5-GCCG- supE44 thi-1 xyl-5 (Str )) (26) was used for cloning purposes and rou- CAAGCTTAGTCCTGCTCCTCGGCC-3) (NcoI and HindIII sites are tinely transformed by electroporation. underlined). PCR products were cloned into vector pET26b (Novagen) Generation of a Bleomycin-based Shuttle Vector—Bacillus subtilis and fused to an Erwinia carotovora pectate lyase (pelB) signal sequence 168 8G5 carrying pUB110 was kindly provided by Dr. S. Bron (Univer- allowing periplasmic protein overexpression in E. coli BL21(DE3) (No- sity of Groningen, Groningen, The Netherlands), and the plasmid was vagen). Periplasmic fractions of 1-liter cultures were prepared by os- isolated by Qiagen Miniprep according to the manufacturer’s instruc- motic shock according to the manufacturer’s instructions and dialyzed tions. The blmS gene was PCR-amplified with primers BG1407 (sense, overnight against 20 mM Tris-HCl (pH 7.5). Samples were loaded onto 5-GGAGGTGCATATGAGAATGTTACAGTCTATCCC-3) and BG1240 a MonoQ HR 5/50 column connected to a fast protein liquid chromatog- (antisense, 5-CGCGTCTAGATTAGCTTTTTATTTGTTGAAAAAAG- raphy system (Amersham Biosciences) and eluted using a 1 M NaCl 3) (NdeI and XbaI sites are underlined). Chromosomal DNA of S. gradient. Shble-containing fractions were pooled and dialyzed against a hindustanus (ATCC 31158) was prepared according to standard proce- 10 mM NaP buffer (pH 7.0) supplemented with 50 mM NaCl and dures (27) and used for PCR amplification of the shble gene with subsequently purified by size exclusion chromatography using a Super- primers BG1410 (sense, 5-TGAGGCATATGGCCAAGTTGACCAGT- dex 200 HR 10/30 column (Amersham Biosciences). GCCG-3) and BG1411 (antisense, 5-GATCCTCTAGATTAGTCCT- Synthetic Gene Construction—A synthetic mutant shble gene based GCTCCTCGGCCACG-3) (NdeI and XbaI sites are underlined). PCR on archaeal codon usage was constructed by oligonucleotide assembly products were digested and ligated into E. coli-T. thermophilus shuttle PCR (31). This gene contains the point mutations G18E, D32V, L63Q, vector pMK18 (28) (Biotools, Madrid, Spain) thereby replacing the and G98V and has a GC content of 40.8% compared with 70.2% of wild-type shble. The synthetic gene was designated HTS (high temper- TM ature Shble). The sequence has been deposited in GenBank under The abbreviations used are: BBP, bleomycin-binding protein; Shble, accession number AY780486. BBP from S. hindustanus; BlmA, BBP from S. verticillus; BlmT, BBP Assembly PCR mixtures contained 10 oligonucleotides (BG1542– from transposon Tn5; BlmS, BBP from plasmid pUB110; HTS, high BG1451, Supplemental Table 1) with an overlap of 20 bases. Both temperature Shble; r.m.s.d., root mean square deviation; DSC, differ- ential scanning calorimetry. flanking primers were 40 bases in length, whereas the eight central 11424 Crystal Structure of Thermostable Bleomycin-binding Protein primers consisted of 80–90 bases. The assembly PCR mixture contained protein was extensively dialyzed against 10 mM NaP , pH 7.0, and was a 2.5 M concentration of each primer, 0.2 mM dNTPs, and 0.05 units/l subsequently crystallized using the sitting drop method of vapor diffu- Pfu polymerase. The mixture was thermocycled at 94 °C (30 s), 55 °C sion at 20 °C and a protein concentration of 3.3 mg/ml in the presence (30 s), and 72 °C (60 s) for 40 cycles. The PCR products were purified of a 10-fold molar excess of bleomycin A2 HCl (Calbiochem). Crystals over a PCR purification column (Qiagen) and diluted 1:1 in fresh PCR grew optimally using 2.0 M ammonium sulfate as the precipitant in 0.1 mixture containing only both flanking primers BG1542 and BG1551 at M sodium acetate buffer, pH 4.6. Data were collected from a single flash 0.1 M concentration and were thermocycled according to standard frozen native crystal (100 K) to 1.5-Å resolution using a MAR345 procedures. PCR products of the expected size were isolated from an imaging plate at the Protein Structure Factory beamline BL14.2 of the agarose gel using the Qiaex II gel extraction kit (Qiagen), digested with Free University of Berlin at the BESSY synchrotron source (Berlin, NdeI and BglII, and cloned into vector pET26b. This allowed for effi- Germany). All data were reduced with DENZO and SCALEPACK (34). cient cytoplasmic protein overproduction in E. coli BL21(DE3)-RIL The crystal used for data collection had unit cell parameters of a  44.0 (Novagen). Positive clones were picked from LB agar plates containing Å, b  66.6 Å, and c  47.2 Å and   117.4° and belonged to space 3 g/ml bleomycin and 50 g/ml chloramphenicol. A 4-liter culture was group P2 with a dimer in the asymmetric unit. grown at 37 °C to an A of 0.5, induced with 0.5 mM isopropyl--D- The structure of HTS was determined by molecular replacement thiogalactopyranoside, and incubated for another 5 h. Cells were har- using the program MOLREP (35) and the S. verticillus BlmA dimer vested, resuspended in 20 mM Tris-HCl (pH 7.5), and sonicated. Cell (Protein Data Bank code 1JIE) (23) as the search model. The initial extracts were clarified by centrifugation (30 min at 26,500  g at 4 °C) phases were improved using the free atom refinement method together and applied to a 70-ml Q-Sepharose Fast Flow (Amersham Biosciences) with automatic model tracing in ARP/wARP (36). Translation, libra- anion exchange column. Proteins were eluted by a 1 M NaCl gradient, tion, and screw rotation (TLS) parameters were determined, and TLS- and HTS-containing fractions were pooled and concentrated over a restrained refinement was performed using REFMAC (37). Several YM10 filter (Amicon) and further purified by size exclusion chromatog- rounds of iterative model building and refinement followed, and water raphy as described above. molecules were added using ARP/wARP (36). The final model (compris- DNA Sequencing—Inserts of plasmids used in this study were se- ing 241 amino acids, 288 water molecules, two bleomycin molecules, quenced by Westburg Genomics (Wageningen, The Netherlands). and three sulfate ions), refined using data between 30- and 1.5-Å Protein Quantitation—Protein concentrations were determined by resolution, has an R- and free R-factor of 17.4 and 19.4%, respectively, using a Bradford assay (32) (Bio-Rad). Purified proteins were quantified with good geometry. Residues Met-1, Glu-122, Gln-123, and Asp-124 in from A measurements using a protein extinction coefficient of 29,000 chain A and Met-1, Gln-123, and Asp-124 in chain B are not visible in 1 1 M cm (33). the electron density map and therefore have been excluded from the DNA Protection Assay—Protein functionality assays were essentially model. Additionally the side chains of two residues in chain A (Asp-36 performed as described elsewhere (14) using a 10-fold excess molar and Arg-87), four side chains in chain B (Glu-21, Asp-36, Arg-87, and concentration of protein over bleomycin A2 (Calbiochem). In each assay, Glu-122), and the -aminopropyldimethylsulfonium moiety of bleomy- 0.2 g of PstI-linearized plasmid, pUC19, was used. Assays were per- cin have been truncated in the final model. The stereochemical quality formed by first incubating DNA and protein shortly at 85 °C after which of the model and the model fit to the diffraction data were analyzed with bleomycin A2, dithiothreitol, and Fe were sequentially added to the the programs PROCHECK (38) and SFCHECK (39). reaction mixture. The coordinates and experimental structure factors have been depos- Circular Dichroism Spectroscopy—CD experiments were performed ited in the Protein Data Bank under accession code 1XRK. Figures were on a Jasco J-715 spectropolarimeter (Jasco, Tokyo, Japan) equipped prepared with Swiss-PDBviewer version 3.7 SP5 (40) and rendered with a PTC-348WI Peltier temperature control system. Far-UV CD with POV-Ray version 3.6. measurements were conducted with Suprasil quartz cuvettes (Hellma Benelux, Rijswijk, The Netherlands) with a 1-mm cell length. During RESULTS AND DISCUSSION all experiments, the sample cell chamber was purged by dry N gas at Selection of Stabilized Variants of Shble a flow rate of 10 liters/min. In temperature-induced unfolding experi- ments, the cuvette containing a 1.7 M concentration of protein sample Randomly mutated shble genes were introduced in the in degassed 10 mM NaP (pH 7.0) and 50 mM NaCl was heated from 25 i E. coli-T. thermophilus shuttle vector pMK18 under the control to 95 °C at 0.4 °C/min and subsequently cooled to 25 °C at the same of the promoter of the surface layer protein A (slpA) from rate. The ellipticity at 205 nm was measured every 0.5 °C with a 2-s T. thermophilus HB8 (41). This promoter is known to drive response time to monitor the loss of -sheet and -turn secondary efficient transcription of the single selection marker in both structure elements. The bandwidth of the measurement was set to 1.0 bacteria. An error-prone library of 20,000 functional clones nm, and the sensitivity was set to 100 millidegrees. Data were corrected for the temperature-dependent ellipticity of a blank without protein. was generated in E. coli HB101. Colonies appeared of similar Averaged data of two independent scans were fit according to a two- size, and there was no difference between the mutant and state model of unfolding, and the apparent temperature unfolding mid- wild-type shble phenotype. The plasmid library was harvested point (T ) was derived from van ’t Hoff plots. and transformed into T. thermophilus HB27, making use of its Fluorescence Spectroscopy—Fluorescence experiments were per- high natural competence (30). Thermus clones appeared on formed on a Varian Cary Eclipse fluorometer (Varian, Middelburg, The bleomycin-containing plates up to 65 °C after transformation Netherlands) equipped with a four-cuvette Peltier multicell holder and PCB-150 waterbath. All measurements were performed in 3-ml Supra- with the wild-type shble shuttle vector, whereas wild-type sil quartz cuvettes (Hellma Benelux) with a 1-cm path length. A mag- blmS was unable to generate a resistant phenotype at either 50 netic stirring bar ensured a homogeneous sample temperature. The or 65 °C. The transformation efficiency of the shble shuttle temperature of the sample was recorded by a temperature probe inside vector was approximately 5 times lower at 65 °C than the one of the four samples. Spectra and thermal unfolding curves were kanamycin-based vector pMK18 (28). This difference might be recorded of 1.7 M protein solutions in degassed 10 mM NaP (pH 7.0) due to the lethal effect of bleomycin and the non-catalytic buffer supplemented with 50 mM NaCl. Tryptophans were excited at 295 nm, and fluorescence was recorded from 300 to 550 nm with both nature of its elimination, which requires at least one protein the excitation and the emission slits set to 5 nm. During temperature- molecule per bleomycin molecule. induced unfolding and refolding studies, fluorescence emission intensi- Upon increasing the temperature of selection, a dramatic ties were monitored at 315 nm from 27 to 92 °C at a heating and cooling decrease in the number of colonies was observed after trans- rate of 0.4 °C/min. Data were corrected for the fluorescence emission of formation of 8 g of mutant library DNA. While 1200 colonies corresponding blank solutions. Data of two independent scans were appeared at 67 °C, this number decreased to 800 at 69 °C, 600 treated and fit as described above. Differential Scanning Calorimetry (DSC)—DSC measurements were at 70 °C, 106 at 75 °C, and eight at 77 °C. No colonies appeared performed on a Microcal III system (Setaram, Caluire, France). De- at 78 and 80 °C. Plating efficiencies at these temperatures have gassed protein samples of 0.28 mg/ml (20 M)in10mM NaP (pH 7.0) been reported to be severely reduced, complicating selection up and 50 mM NaCl in the presence and absence of an 8-fold molar excess to 85 °C, the maximum temperature of growth (11). The eight of bleomycin A2 were heated from 20 to 120 °C at 0.5 °C/min. Midpoint Thermus clones found at 77 °C (termed 77-1 to 77-8) were temperatures of unfolding were determined by curved base-line analy- grown overnight in selective medium at 70 °C; their plasmids sis from two independent scans. Protein Crystallization, Data Collection, and Processing—The HTS were isolated, transformed into E. coli HB101, and subse- Crystal Structure of Thermostable Bleomycin-binding Protein 11425 TABLE I double mutants, only 77-3 (D32V,L63Q) seems to have a Nucleotide and amino acid substitutions of Shble variants marked increase in T as observed with CD spectroscopy and Residue number fluorescence spectroscopy, while numbers 77-1 (L63Q,G98V), 77-5 (R31L,G98S), and 77-7 (G18E,L63Q) remain virtually un- 18 31 32 40 63 98 changed. Surprisingly mutant 77-4 (R31L,G40A) displays sig- Wild type Gly Arg Asp Gly Leu Gly nificantly lower T values compared with the wild-type. Quad- Gln Val 77-1/77-2/77-8 T188A G293T ruple mutant HTS, which combines non-redundant mutations Val Gln found in 77-1, 77-3, and 77-7, displays a profound increase of 77-3 A95T T188A 13.9, 10.8, and 17.7 °C in stability in the absence of the anti- Leu Ala 77-4/77-6 biotic as found by CD, fluorescence spectroscopy, and DSC, G92T G119C Leu Ser respectively. In its presence, the complex becomes hyperther- 77-5 G92T G292A mostable, unfolding at a temperature of just over 100 °C, 5.6 °C Glu Gln 77-7 higher compared with the wild-type protein. G53A T188A To our surprise, the double mutants 77-1, 77-5, and 77-7 had Clone contained silent nucleotide substitution C360A. almost unchanged apparent melting temperatures compared Clone contained silent nucleotide substitution G63A. c with the wild-type. This can be understood by realizing that in Clone contained silent nucleotide substitution G30A. vivo some amino acid changes may prevent instances of local protein unfolding and therefore may avoid further unfolding quently reisolated; and their inserts were sequenced. This re- and subsequent proteolytic attack. However, this is not neces- vealed that all variants were double mutants bearing, in total, sarily reflected in its in vitro melting temperature, which is a six different amino acid substitutions and three silent muta- measure of its global stability. Only when the weakest point of tions (Table I). Five types of double mutants could be distin- a structure was compensated (D32V and L63Q in 77-3), an guished at the protein level, and two sets of double mutants increase of its melting temperature from 70.8 to 79.5 °C with were identical. Remarkably three of six mutations found were CD and 67.9 to 69.1 °C with fluorescence spectroscopy was glycine substitutions of which glycine 98 was replaced by either observed. Adding mutation G98V from 77-1 and G18E from a valine or a serine. The fact that only double mutants were 77-7 to 77-3, giving rise to HTS, further increased its melting found seems to be a clear indication of the high stringency that temperature as one would expect. This observation is analo- was used during selection. Interestingly some substitutions, gous to the findings of extensive work that has been conducted such as L63Q, had occurred in combination with either G18E, with the neutral protease from G. stearothermophilus where D32V, or G98V, which may point to the independent effects of interactions close to the N terminus were found to be limiting the different mutations. A multiple sequence alignment of the global stability (5). BBPs and the position of the mutations are shown in Fig. 1. To assess the reason why these mutants performed better at ele- Mutants Improve DNA Protection against Bleomycin at vated temperatures in vivo, we produced and purified wild-type High Temperature Shble and all double mutants and studied their biochemical In vitro DNA protection assays were performed with the behavior in vitro. Furthermore a synthetic quadruple mutant gene with low GC content was designed by combining muta- various Shble mutants to test whether the resistant phenotype of T. thermophilus at 77 °C was due to improved protection tions G18E, D32V, L63Q, and G98V. The protein, designated HTS, was produced, purified, and biochemically analyzed. The against the DNA degrading capability of bleomycin. The result of this is shown in Fig. 2. At 25 °C, no significant differences in HTS protein was crystallized in complex with bleomycin A2, and its structure was determined. band intensities were observed. A 30-min thermal preincuba- tion of the protein at 85 °C, however, revealed a drastic loss of Thermal Unfolding function in mutant 77-4. Differences between the wild-type and mutants became pronounced when bleomycin binding capabil- Shble variants were subjected to temperature-induced equi- ities were tested at 85 °C. At this temperature, the DNA was librium unfolding experiments in the presence and absence of protected best by 77-1 and HTS followed by 77-4, 77-5, 77-3, bleomycin. The protein was found to unfold largely irreversibly 77-7, and the wild type. Surprisingly mutant 77-4, which dis- since only 40% of the native folded signal was regained after played a low temperature unfolding midpoint and high thermal slow cooling of the thermally unfolded protein. Therefore only inactivation at 85 °C, apparently bound bleomycin effectively apparent midpoint temperatures of unfolding (T ) could be at high temperature conditions. So although the global stability calculated. The results are summarized in Table II. of this mutant was decreased, it had improved bleomycin bind- In the absence of the antibiotic, wild-type Shble appears to be ing characteristics, which in itself stabilizes the protein dra- a very stable protein. This is remarkable because S. hindusta- matically as observed by DSC measurements for the wild type. nus grows optimally at 28 °C (42). It is often found, however, These results indicate that some of the double mutants have that proteins for which low biological turnover is beneficial for improved the bleomycin binding properties compared with the a host are prone to little local unfolding and hence are less wild type, confirming the findings of the in vivo selection pro- susceptible to proteolytic attack (43). Structurally Shble, which cedure in T. thermophilus. Possible structural explanations for serves a function of self-immunity, might well be adapted to the improved functionality at higher temperature are dis- meet these criteria by its compactness, relatively high second- cussed below. ary structure content, high surface charge, and embedded N and C termini (19). The unfolding data also clearly show the Overall Structure Description strong stabilizing effect of ligand binding on the thermostabil- ity of the BBP since the apparent unfolding midpoint temper- The quadruple mutant HTS was crystallized in the presence ature increases 27.3 °C upon bleomycin binding. This effect has of bleomycin A2, and its structure was determined to 1.50-Å also been recognized in other ligand-binding proteins, such as resolution (Table III). The crystals grown belong to space group streptavidin and avidin, which become extremely thermostable P2 with unit cell parameters of a  44.0 Å, b  66.6 Å, and c in the presence of biotin (44). In the absence of bleomycin, the 47.2 Å and   117.4° and a dimer in the asymmetric unit (Fig. stability of the various mutants is rather different. Of the 3, A and B). Representative electron density is shown in Fig. 11426 Crystal Structure of Thermostable Bleomycin-binding Protein FIG.1. Structural alignment of bleomycin-binding proteins from different microbial sources. Shble (Protein Data Bank code 1BYL) from S. hindustanus (19), BlmA (Protein Data Bank code 1QTO) from S. verticillus ATCC15003 (45), SvP from S. verticillus ATCC21890, BlmT (Protein Data Bank code 1ECS) from K. pneumoniae transposon Tn5 (22), and BlmS from S. aureus plasmid pUB110 are shown. The alignment was created by backbone superimposition of the three structures and expanded with the SvP and BlmS sequences by realignment using ClustalX version 1.81 (65) while maintaining the original gaps. The HTS structure was used for residue numbering and topology assignment (black arrows, -strand; checkered boxes, -helix). Mutations are indicated by arrows. TABLE II Apparent thermal unfolding midpoints (°C) of Shble variants WT HTS 77-1 77-3 77-4 77-5 77-7 Uncomplexed CD 70.8  0.5 84.7  0.6 72.2  0.5 79.5  0.6 66.8  0.5 71.1  0.6 70.3  0.5 FS 67.9  0.4 78.7  0.5 67.2  0.4 69.1  0.5 63.8  0.4 64.6  0.5 65.9  0.4 DSC 67.4  0.6 85.1  0.7 ND ND ND ND ND Bleomycin-bound DSC 94.7  0.8 100.3  0.8 ND ND ND ND ND Wild type. Circular dichroism spectroscopy ( ). 205 nm Fluorescence spectroscopy ( ,  ). ex 295 nm em 315 nm ND, not determined. TABLE III Data collection and refinement statistics Values in parentheses refer to the highest resolution shell. X-ray data collection statistics Wavelength (Å) 0.90830 Resolution (Å) 30–1.5 (1.53–1.50) Total observations 87,740 (3,602) Unique observations 36,444 (1,714) Completeness (%) 94.4 (88.6) I/(I) 20.3 (3.9) R 0.043 (0.189) sym Refinement statistics Resolution (Å) 30.0–1.5 R 0.174 FIG.2. DNA protection assay. Digital photographs of 1% agarose work R 0.194 free gels showing the degree of DNA protection by Shble variants against r.m.s.d. bond distances (Å) 0.01 the strand scission action of bleomycin A2 are shown. A, assay at 25 °C r.m.s.d. bond angles (°) 1.664 for 10 min. B, assay at 25 °C for 10 min after protein preincubation at Total number of non-hydrogen atoms 1,883 85 °C for 30 min. C, assay at 85 °C for 10 min. ble, bleomycin A2; WT, Average protein B-value (Å ) 13.3 wild type. Number of solvent molecules 288 Average solvent B-value (Å ) 27.0 3C. The structure forms a compact, homodimeric / protein of Number of bleomycin atoms 182 Average ligand B-value (Å ) 22.8 121 amino acids (Met-1, Gln-123, and Asp-124 are disordered) Number of sulfate ion atoms 15 in which two bleomycin A2 molecules are accommodated in Average sulfate ion B-value (Å ) 31.4 binding pockets at the dimer interface. These pockets consist of R  I  I / I where I is the intensity of a given sym hkl i i hkl i i i a hydrophilic concavity that runs into a hydrophobic intersub- measurement, and the sums are over all measurements and reflections. unit crevice. The dimer is maintained by alternate N-terminal b R  F   F / F  for 95% of the reflection data used in work obs calc obs -strand hydrogen bonding between both monomers and by refinement. R  F   F / F  for the remaining 5%. Van der Waals interactions at the largely hydrophobic subunit free obs calc obs contact (19, 45). Three sulfate ions are present at the surface of the dimer of which two form ion pairs with Arg-104 of both (r.m.s.d., 0.38 Å; Table IV) only revealed large differences in a chains. The presence of a dimer in the asymmetric unit allowed random coil region comprising residues Asp-88, Ala-89, and the identification of certain symmetry deviations between both Ser-90 (Fig. 3, A and B) that is spatially close to the carbamoyl monomers. A backbone superimposition of both chains group of the D-mannose moiety of bleomycin (Fig. 4A). Their Crystal Structure of Thermostable Bleomycin-binding Protein 11427 FIG.3. Structure and electron density of HTS in complex with bleomycin A2. A ribbon diagram showing the dimeric structure of the 4-fold mutant Shble in complex with bleomycin A2 is shown. Mutations are indicated by stick representations. Chain A is in blue, chain B is in red, G18E is in green, D32V is in pink, L63Q is in yellow, and G98V is in orange. A, side view. B, viewed from the N- and C-terminal side (top view). C, stereoview of the electron density around residues Pro-9 and Trp-65 contoured at 2 . Residues are colored according to the Corey-Pauling- Koltun color scheme, and water molecules are represented by red spheres. TABLE IV Crystal structures of bleomycin-binding proteins Protein Data Protein/source Form Resolution Backbone r.m.s.d. Ref. Bank code ÅÅ 1BYL Uncomplexed 2.30 0.63 19 Shble S. hindustanus 1XRK Bleomycin A2-bound 1.50 0.38 This study BlmA S. verticillus 1QTO Uncomplexed 1.50 0.77 45 1JIE Bleomycin A2-bound 1.80 0.73 23 2 a 1JIF Cu -bleomycin A2-bound 1.60 0.75 23 1ECS Uncomplexed 1.70 1.18 22 BlmT K. pneumoniae transposon Tn5 1EWJ Bleomycin A2-bound 2.50 1.15 22 Averaged backbone superimposition r.m.s.d. values compared with HTS chains A and B. Backbone superimposition r.m.s.d. values of HTS chain A to B. respective C atoms deviate 2.0, 5.1, and 1.6 Å in position while standing of these proteins in general. Moreover the structure giving rise to almost oppositely pointing amino acid side has revealed several molecular features that can account for chains. In contrast to the bleomycin-bound and unbound BlmA increased protein stability and improved functionality at structure, backbone B-factors in this region are only margin- higher temperature in vivo and in vitro. ally higher compared with the average value, suggesting a rigid Structural Effects of Mutations conformation (23, 45). The difference in orientation of this loop might therefore be the result of sequential binding of two Introduction of an Intersubunit Hydrogen Bond Network— bleomycin molecules. Unlike BlmA, no symmetry-related dif- The structure of the dimer shows that each of the two bleomycin ferences were observed in the region between amino acids 100 A2 molecules is bound by the concerted action of 21 amino acids. and 103. The topology of the HTS protein complex and the Due to its intersubunit location, both binding sites are composed mode of bleomycin binding are similar to other BBPs. An over- of residues from either subunit. These include Val-32, Phe-33, view of available structures is given in Table IV. Glu-35, Phe-38, Ser-51, Ala-52, and Val-53 of one subunit and The structure of the HTS mutant in complex with bleomycin Pro-59, Asp-60, Asn-61, Thr-62, Gln-63, Trp-65, Phe-86, Ala-89, completes the list of structural information of three BBPs with Trp-102, Ala-107, Arg-109, Gly-113, Cys-115, and His-117 of the and without their ligands, hereby contributing to our under- other. The crystal structure clearly reveals the central role of 11428 Crystal Structure of Thermostable Bleomycin-binding Protein FIG.4. Chemical diagram and electron density of bleomycin A2. A, schematic representation of bleomycin A2. B, electron density around bleomycin A2 contoured at 1.5 . The diagram indicates the missing electron density around the -aminopropyldimethylsulfonium moiety suggesting a disordered conformation. mutation L63Q, which was found in three of five different double ing hydrophobic packing of surface indentations, thereby rein- mutants. Gln-63 is involved in an extensive hydrogen bond net- forcing subsurface secondary structure elements (48). This work at the bottom of the bleomycin binding concavity (Fig. 5A). might explain the enhanced secondary structure preservation It is noteworthy that the carbonyl side chains (O-1) of both at high temperatures as inferred from CD of mutant 77-3. Gln-63 residues in the dimer act as terminal hydrogen bond Strikingly modeling of G40A into the wild-type structure (not acceptors of a five-molecule water channel present at the dimer shown) revealed close spatial proximity to Asp-32 (5 Å between interface. A second hydrogen bond is accepted from the side chain C and 3.5 Å between C atoms) that may also underline a hydroxyl group (O-) of Ser-51 of the adjacent subunit. The amide similar need for hydrophobicity in this part of the protein. side chain (N-2) of Gln-63 forms a hydrogen bond with one of two Mutation Arg-31 to Leu, which occurred in two types of double water molecules trapped between the bleomycin and the surface mutants, came as a surprise since it is involved in a surface ion of the protein. The presence of a leucine at position 63 would most pair with Asp-25 in wild-type Shble. Apparently this electro- likely not have allowed for a hydrogen bond network of this size. static interaction does not counterweight the beneficial effects The advantage of an amino acid compatible with hydrogen bond- of improved hydrophobic packing among residues Val-20, Thr- ing at position 63 is also evident from the alignment, which 24, Val-34, and Val-41. indicates that without exception the other four BBPs have a Reduction of Surface Loop Flexibility—From previous crys- serine at this specific site (Fig. 1). Although the mutant structure tallographic and NMR studies of BBPs, it has become clear that without bleomycin is not available, we speculate that an inter- the loop following Gly-98 in Shble will change its conformation subunit hydrogen bond between Gln-63 and Ser-51 can persist upon binding of the antibiotic (22, 23, 45, 46). This conforma- even without the antibiotic bound, giving rise to a beneficial tional change enables the tryptophan at position 102 to stack interaction that might stabilize the dimer at high temperatures. optimally with the hydrophobic bithiazole moiety of bleomycin Hydrophobic Packing of Surface Indentations—In wild-type (Fig. 4A), packing both thiazole rings tightly against Phe-33 Shble and BlmA, Asp-32 is located on the edge of the intersub- and Phe-38 of the adjacent subunit. In both BlmA and Shble, unit binding groove for the bithiazole moiety and tail region of Gly-98, located on the edge of a small -strand leading toward bleomycin (Fig. 4A) (19, 23, 45). In bleomycin A2 and B2 and in the binding loop, seems to have a hinge function (Fig. 5C). The phleomycin D1 the tail is positively charged, suggesting in- bending motion of the backbone is also clearly reflected in large volvement of Asp-32 in electrostatic stabilization or ligand  and torsion angle changes of more than 20° upon the recognition. From NMR studies it has become clear that in the binding of bleomycin. At high temperatures, however, this flex- bound state no strong interactions occur between the protein ibility might have caused problems leading to local unfolding or and the positively charged tail of bleomycin (46). These data a decreased bleomycin binding ability. A substitution for either are supported by the absence of electron density for the -amin- a valine or serine as observed would increase the rigidity of the opropyldimethylsulfonium moiety of bleomycin A2 in the bi- loop and could therefore restore the binding capacity at nary complex structure of BlmA and HTS, suggesting a disor- high temperature. dered conformation of the tail end (Fig. 4B) (23). This might -Helix Stabilization—Mutation G18E introduces a gluta- have allowed for an amino acid substitution to valine, which mate at position N-3 in the first turn of the largest -helix of the extends the hydrophobic bithiazole binding cleft at the dimer protein. Statistical analysis and experimental studies have interface fitting nicely within a highly hydrophobic environ- shown that glutamates are energetically highly favored over ment consisting of Phe-33, Phe-38, Val-42, Thr-47, and Phe-49 glycines at the third position in an -helix (49, 50). This effect is (Fig. 5B). In addition, both BlmT and BlmS sequences also most likely caused by the stabilizing effect of the negatively contain a valine at the corresponding position (Fig. 1). From a charged side chain on the helix macrodipole. To our surprise, thermodynamic point of view, a mutation introducing surface chain A of the crystal structure revealed the formation of a hydrophobicity is generally believed to be unfavorable and has genuine i, i  5 -helix surface ion pair between Glu-21 and therefore rarely been investigated in directed mutagenesis Arg-26 that was absent in the wild-type structure (Fig. 5D). This studies. Nevertheless some studies have reported significant new ion pair may have been the result of repulsion of anionic improvements in protein stability by placing bulky hydropho- glutamate side chains of positions 18 and 21, directing the latter bic amino acids at the surface of a neutral protease from toward the C-terminal arginine. Although i, i  5 -helical sur- G. stearothermophilus (47). Recent findings using Bacillus li- face ion pairs do not give rise to strong ionic interactions at cheniformis -amylase have clearly indicated that hydrophobic ambient temperatures (51), they might be more favorable at surface residues can indeed be extremely stabilizing by improv- higher temperatures. Theoretical models have indicated that the Crystal Structure of Thermostable Bleomycin-binding Protein 11429 FIG.5. Structural effects of the individual mutations. A, L63Q. A ribbon diagram showing the hydrogen bond network at the dimer interface is shown. Thr-62, Gln-63, Ser-51, and bleomycin A2 are shown together with the intersubunit water channel. B, D32V. A ribbon diagram showing the hydrophobic intersubunit bleomycin tail binding crevice is shown. Val-32 may be involved in improved hydrophobic packing of this surface indentation among amino acids Phe-33, Phe-38, Val-42, Thr-47, and Phe-49. C, G98V. A ribbon representation showing a loop between Pro-92 and Pro-111 that is involved in bleomycin binding is shown. Val-98 is located at a former hinge region that enables Trp-102 to stack the bithiazole tail against Phe-33 and Phe-38. The electron density revealed two alternative side chain rotamers for Val-98 in chain A (not shown) and a single side chain conformation in chain B. D, G18E. A side-by-side comparison of -helix 1 formed between Asp-15 and Leu-27 in the wild-type and mutant crystal structures of Shble. The existence of a surface ion pair between Glu-21 and Arg-26 is visible in the electron density (contoured at 1.5 ). Interatomic distances are indicated in Å. Corey-Pauling-Koltun color coding was used for amino acids and bleomycin A2. Chain A is indicated in blue, and chain B is indicated in red. Water molecules are represented by red spheres. Hydrogen bonds are depicted by green dotted lines. energetic cost of desolvating charged groups is much less at residues (Asn, Gln, Ser, and Thr) (54). This is fully in agree- 100 °C due to a drop in the dielectric constant of water (52). This ment with the requirements for the elevated numbers of sur- is currently the best explanation for the fact that proteins from face salt bridges and improved hydrophobic core packing that extreme thermophiles have large ion pair networks at their sur- has generally been recognized in these types of proteins. These faces that are thought to be involved in maintaining structural are just two of a multitude of mechanisms that proteins from integrity (3). Additionally a minor beneficial effect of this muta- hyperthermophilic microorganisms have used to deal with ex- tion could be the introduction of additional negative surface treme temperatures (2, 55, 56). charge, which enhances electrostatic attraction of the cationic Recently several other random mutagenesis studies have antibiotic under physiological conditions. also reported large improvements in thermostability by apply- Laboratory Versus Natural Evolution of Thermostability—In ing directed evolution approaches. A mesophilic xylanase of this study, several possible mechanisms of adaptation to high family 11 was stabilized by over 35 °C by combining nine mu- temperature were identified, such as the introduction of a tations found separately after extensive screening. The activity hydrogen bond network, improved hydrophobic packing of sur- of this mutant was optimized by saturation mutagenesis of all face indentations, reduction of loop flexibility, and -helix sta- mutated positions, yielding an enzyme variant with highly bilization. Remarkably half of all mutations found were glycine enhanced properties for high temperature applications (8). In replacements, which could point to protein stabilization by another study, a highly thermostable esterase containing seven decreasing the entropy of the unfolded state (53). Although this mutations was evolved in six rounds of random mutagenesis, could be a general strategy of stabilization, proteins from hy- recombination, and screening (7). The resulting enzyme was perthermophiles do not have a lower glycine content than their crystallized, and its structure was determined (57). The struc- mesophilic counterparts but rather a slightly increased one ture revealed that improved stability was due to altered core (54). Their predicted proteomes do have an increased propen- packing, -helix stabilization, the introduction of surface salt sity for charged (Arg, Lys, and Glu) and bulky aliphatic (Ile and bridges, and reduction of flexibility in surface loops. From these Val) amino acids that has mostly come at the cost of polar and many other directed evolution and site-directed mutagen- 11430 Crystal Structure of Thermostable Bleomycin-binding Protein Acknowledgments—We thank Dr. J. Berenguer for helpful sugges- esis studies, it has become apparent that (i) proteins can be tions and Anton Korteweg for technical assistance with DSC. stabilized substantially by small numbers of mutations, (ii) these mutations are often located at the protein surface, and REFERENCES (iii) their effects are usually additive. As few as two of 12 amino 1. Arnold, F. H., Wintrode, P. L., Miyazaki, K., and Gershenson, A. (2001) Trends acid differences between a mesophilic and thermophilic cold Biochem. Sci. 26, 100–106 2. Vieille, C., and Zeikus, G. J. (2001) Microbiol. Mol. Biol. Rev. 65, 1–43 shock protein turned out to be responsible for the difference in 3. Karshikoff, A., and Ladenstein, R. (2001) Trends Biochem. Sci. 26, 550–556 thermostability (58). The remaining variation in sequence 4. Serrano, L., and Fersht, A. R. (1989) Nature 342, 296–299 might just have occurred as a result of neutral sequence drift or 5. Van den Burg, B., Vriend, G., Veltman, O. R., Venema, G., and Eijsink, V. G. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 2056–2060 specific properties required by the host, such as solubility, 6. Stemmer, W. P. (1994) Nature 370, 389–391 turnover, and molecular interactions (59). 7. Giver, L., Gershenson, A., Freskgard, P. O., and Arnold, F. H. (1998) Proc. Despite the fact that only a small number of mutations are Natl. Acad. Sci. U. S. A. 95, 12809–12813 8. Palackal, N., Brennan, Y., Callen, W. N., Dupree, P., Frey, G., Goubet, F., required to render a protein thermostable, finding those muta- Hazlewood, G. P., Healey, S., Kang, Y. E., Kretz, K. A., Lee, E., Tan, X., tions remains a difficult task. Apart from screening vast num- Tomlinson, G. L., Verruto, J., Wong, V. W., Mathur, E. J., Short, J. M., Robertson, D. E., and Steer, B. A. (2004) Protein Sci. 13, 494–503 bers of random mutants in microtiter plates, in vitro and in 9. Liao, H., McKenzie, T., and Hageman, R. (1986) Proc. Natl. Acad. Sci. U. S. A. vivo selection schemes offer great advantages to reduce the 83, 576–580 10. Tamakoshi, M., Yamagishi, A., and Oshima, T. (1995) Mol. Microbiol. 16, effort to encounter improved variants. Currently two main 1031–1036 strategies are available. 11. Hoseki, J., Yano, T., Koyama, Y., Kuramitsu, S., and Kagamiyama, H. (1999) One of these thermostabilizing selection methods is called J. Biochem. (Tokyo) 126, 951–956 12. Friedrich, A., Prust, C., Hartsch, T., Henne, A., and Averhoff, B. (2002) Appl. Proside (protein stability increased by directed evolution) Environ. Microbiol. 68, 745–755 which is based on the empirically derived inverse correlation 13. Noll, K. M., and Vargas, M. (1997) Arch. Microbiol. 168, 73–80 between protein thermostability and proteolytic susceptibility 14. Gatignol, A., Durand, H., and Tiraby, G. (1988) FEBS Lett. 230, 171–175 15. Sugiyama, M., and Kumagai, T. (2002) J. Biosci. Bioeng. 93, 105–116 (60). In a phage display-like procedure, a library of mutants is 16. Poddevin, B., Orlowski, S., Belehradek, J., Jr., and Mir, L. M. (1991) Biochem. fused between two domains of E. coli filamentous phage Fd Pharmacol. 42, (suppl.) S67–S75 17. Sikic, B. I. (1986) Cancer Surv. 5, 81–91 gene-3-protein, which is then subjected to proteases while in- 18. Sugiyama, M., Thompson, C. J., Kumagai, T., Suzuki, K., Deblaere, R., Villar- ducing local unfolding of the target protein by means of tem- roel, R., and Davies, J. (1994) Gene (Amst.) 151, 11–16 perature or chemical denaturants. The resulting unstable fu- 19. Dumas, P., Bergdoll, M., Cagnon, C., and Masson, J. M. (1994) EMBO J. 13, 2483–2492 sion protein variants are cleaved, causing only the surviving, 20. Genilloud, O., Garrido, M. C., and Moreno, F. (1984) Gene (Amst.) 32, 225–233 more stable phages to be found after infection. It was found 21. Semon, D., Movva, N. R., Smith, T. F., el Alama, M., and Davies, J. (1987) Plasmid 17, 46–53 that in an ionic denaturant, non-polar surface interactions 22. Maruyama, M., Kumagai, T., Matoba, Y., Hayashida, M., Fujii, T., Hata, Y., were optimized, whereas at elevated temperature variants and Sugiyama, M. (2001) J. Biol. Chem. 276, 9992–9999 with improved surface electrostatics were selected (59). 23. Sugiyama, M., Kumagai, T., Hayashida, M., Maruyama, M., and Matoba, Y. (2002) J. Biol. Chem. 277, 2311–2320 Cloning and selection in a thermophile, as conducted in this 24. Drocourt, D., Calmels, T., Reynes, J. P., Baron, M., and Tiraby, G. (1990) study, is a second directed evolution strategy that can lead to Nucleic Acids Res. 18, 4009 25. Nuttall, S. D., Deutschel, S. E., Irving, R. A., Serrano-Gomicia, J. A., and rapid improvements in thermostability. Despite its simplicity, Dyall-Smith, M. L. (2000) Biochem. J. 346, 251–254 very few studies using this technique have been reported. This 26. Boyer, H. W., and Rouland-Dussoix, R. (1969) J. Mol. Biol. 41, 459–472 is most likely due to its limited applicability because the gene 27. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd Ed., Cold Spring Harbor Laboratory, Cold Spring of interest rarely confers any biologically relevant function for Harbor, NY growth or survival of the thermophile. Nonetheless function- 28. de Grado, M., Castan, P., and Berenguer, J. (1999) Plasmid 42, 241–245 29. Shafikhani, S., Siegel, R. A., Ferrari, E., and Schellenberger, V. (1997) Bio- ally stabilized mutants have been reported up to 79 °C for a Techniques 23, 304–310 kanamycin nucleotidyltransferase (9, 11) and 58 °C for a chlor- 30. Koyama, Y., Hoshino, T., Tomizuka, N., and Furukawa, K. (1986) J. Bacteriol. amphenicol acetyltransferase (61). Auxotrophic knock-out 166, 338–340 31. Stemmer, W. P., Crameri, A., Ha, K. D., Brennan, T. M., and Heyneker, H. L. strains for leucine biosynthesis were complemented with ther- (1995) Gene (Amst.) 164, 49–53 molabile counterpart genes from B. subtilis and Saccharomyces 32. Bradford, M. M. (1976) Anal. Biochem. 72, 248–254 33. Rondeau, J. M., Cagnon, C., Moras, D., and Masson, J. M. (1989) J. Mol. Biol. cerevisiae and adapted to higher temperature by serial accu- 207, 645–646 mulation of beneficial mutations in the in trans introduced 34. Otwinowski, A., and Minor, W. (1997) Methods Enzymol. 276, 307–326 3-isopropylmalate dehydrogenase genes (62, 63). A hybrid -ga- 35. Vagin, A. A., and Teplyakov, A. (1997) J. Appl. Crystallogr. 30, 1022–1025 36. Perrakis, A., Morris, R., and Lamzin, V. S. (1999) Nat. Struct. Biol. 6, 458–463 lactosidase consisting of Bacillus stearothermophilus and T. 37. Murshudov, G. N., Vagin, A. A., and Dodson, E. J. (1997) Acta Crystallogr. thermophilus peptide regions was adapted to function at 67 °C Sect. D. Biol. Crystallogr. 53, 240–255 by selection for growth on melibiose as the sole carbon and 38. Laskowski, R. A., MacArthur, M. W., Moss, D. S., and Thornton, J. M. (1993) J. Appl. Crystallogr. 26, 283–291 energy source (64). 39. Vaguine, A. A., Richelle, J., and Wodak, S. J. (1999) Acta Crystallogr. Sect. D Biol. Crystallogr. 55, 191–205 40. Guex, N., and Peitsch, M. C. (1997) Electrophoresis 18, 2714–2723 Concluding Remarks 41. Fernandez-Herrero, L. A., Olabarria, G., and Berenguer, J. (1997) Mol. Micro- biol. 24, 61–72 In this study, several double mutants of a bleomycin-binding 42. Tomita, K., Nakakita, Y., Hoshino, Y., Numata, K., and Kawaguchi, H. (1987) Int. J. Syst. Bacteriol. 37, 211–213 protein were isolated with enhanced performance at high tem- 43. Jaswal, S. S., Sohl, J. L., Davis, J. H., and Agard, D. A. (2002) Nature 415, perature both in vivo and in vitro. Structural analysis showed 343–346 that the mutations gave rise to different means of stabilization 44. Gonzalez, M., Argarana, C. E., and Fidelio, G. D. (1999) Biomol. Eng. 16, 67–72 45. Kawano, Y., Kumagai, T., Muta, K., Matoba, Y., Davies, J., and Sugiyama, M. in four parts of the protein. A combined mutant gene with a low (2000) J. Mol. Biol. 295, 915–925 GC content was created that can serve as an antibiotic resist- 46. Vanbelle, C., Brutscher, B., Blackledge, M., Muhle-Goll, C., Remy, M. H., Masson, J. M., and Marion, D. (2003) Biochemistry 42, 651–663 ance marker for aerobic and microaerophilic mesophiles, ther- 47. Van den Burg, B., Dijkstra, B. W., Vriend, G., Van der Vinne, B., Venema, G., mophiles, and hyperthermophiles. This may allow the develop- and Eijsink, V. G. (1994) Eur. J. Biochem. 220, 981–985 48. Machius, M., Declerck, N., Huber, R., and Wiegand, G. (2003) J. Biol. Chem. ment of efficient shuttle vectors and knock-out strategies for 278, 11546–11553 hyperthermophilic Archaea and bacteria based on positive se- 49. Iqbalsyah, T. M., and Doig, A. J. (2004) Protein Sci. 13, 32–39 lection schemes. Moreover the high GC content double mutant 50. Penel, S., Hughes, E., and Doig, A. J. (1999) J. Mol. Biol. 287, 127–143 51. Smith, J. S., and Scholtz, J. M. (1998) Biochemistry 37, 33–40 genes will now permit multigene knock-out strategies in ther- 52. Elcock, A. H. (1998) J. Mol. Biol. 284, 489–502 mophiles such as T. thermophilus, allowing further exploration 53. Van den Burg, B., and Eijsink, V. G. (2002) Curr. Opin. Biotechnol. 13, and exploitation of thermophilic microbial sources. 333–337 Crystal Structure of Thermostable Bleomycin-binding Protein 11431 54. Cambillau, C., and Claverie, J. M. (2000) J. Biol. Chem. 275, 32383–32386 61. Turner, S. L., Ford, G. C., Mountain, A., and Moir, A. (1992) Protein Eng. 5, 55. Kumar, S., and Nussinov, R. (2001) Cell Mol. Life Sci. 58, 1216–1233 535–541 56. Vogt, G., Woell, S., and Argos, P. (1997) J. Mol. Biol. 269, 631–643 62. Akanuma, S., Yamagishi, A., Tanaka, N., and Oshima, T. (1998) Protein Sci. 7, 57. Spiller, B., Gershenson, A., Arnold, F. H., and Stevens, R. C. (1999) Proc. Natl. 698–705 Acad. Sci. U. S. A. 96, 12305–12310 63. Tamakoshi, M., Nakano, Y., Kakizawa, S., Yamagishi, A., and Oshima, T. 58. Perl, D., Mueller, U., Heinemann, U., and Schmid, F. X. (2000) Nat. Struct. (2001) Extremophiles 5, 17–22 Biol. 7, 380–383 64. Fridjonsson, O., Watzlawick, H., and Mattes, R. (2002) J. Bacteriol. 184, 59. Martin, A., Sieber, V., and Schmid, F. X. (2001) J. Mol. Biol. 309, 717–726 3385–3391 60. Sieber, V., Pluckthun, A., and Schmid, F. X. (1998) Nat. Biotechnol. 16, 65. Thompson, J. D., Gibson, T. J., Plewniak, F., Jeanmougin, F., and Higgins, 955–960 D. G. (1997) Nucleic Acids Res. 25, 4876–4882

Journal

Journal of Biological Chemistry – American Society for Biochemistry and Molecular Biology

Published: Mar 25, 2005

There are no references for this article.