Free energy perturbation study on a Trp-binding mutant (Ser88 ->Cys) of the trp-repressor

The Ser⁸⁸Cys mutant of the trp-repressor showed a lower affinityfor the corepressor than the wild-type repressor [G = 1.7 ±0.3 kcal/mol, Chou and Matthews (1989) J. Biol. Chem., 264,18314–18319].A molecular dynamics/free energy cycle perturbation study wasperformed to understand the origin of the decreased affinity.A value (G = 1.58 ± 0.28 kcal/mol) comparable with theexperimental value was obtained by the simulation. Free energycomponent analysis revealed that destabilization of the vander Waals interaction between Ser⁸⁸ and Trp¹⁰⁹ (corepressor)mainly contributed to the decreased affinity of the mutant.The rotational transition of the hydroxyl (sulfhydryl) groupof Ser⁸⁸ (Cys⁸⁸) during the simulations affected the contributionsof Arg⁸⁴ and water to the free energy change in the aporepressorand those of Arg⁸⁴ and Trp ¹⁰⁹ to that in the holorepressor.However, the contributions from different residues compensatedeach other, and the total free energy changes were almost invariablein the various simulations.     

Effect of amino acid deletions in the O-glycosylated region of Aspergillus awamori glucoamylase

Aspergillus awamori glucoamylase (GA) contains globular catalyticand starch-binding domains (residues 1–471 and 509–616,respectively). A heavily O-glycosylated sequence comprises twoparts. The first (residues 441–471) in the crystal structurewraps around an /-barrel formed by residues 1–440. Thesecond (residues 472–508) is an extended, semi-rigid linkerbetween the two domains. To investigate the functional roleof this linker, we made internal deletions to remove residues466–512 (GA1), 485–512 (GA2) and 466–483 (GA3).GA2 and GA3 were expressed in Saccharomyces cerevisiae culturesupernatants at 60 and 20% the wild-type level, respectively,while GA1 was almost undetectable. Western blots comparing extracellularand intracellular fractions indicated that the region deletedin GA3 was critical for secretion, while the region deletedin GA2 contributed to the production of a stable enzyme structure.The activities of purified GA2 and GA3 on soluble and insolublestarch were similar to those of wild-type GA, indicating thatfor soluble starch their deletions did not affect the catalyticdomain and for insoluble starch the linker does not coordinatethe activities of the catalytic and starch-binding domains.The deletions had a significant negative effect on GA2 and GA3thermos tabilities.     

The net energetic contribution of interhelical electrostatic attractions to coiled-coil stability

The net energetic contribution of interhelical electrostaticattractions to coiled-coil stability has been quantitated usingde novo designed synthetic coiled-coils. The synthesized modelcoiled-coil (EK), denoted by amino acid res-idues in positionse and g, which contains only interhelical ionic interactionswithout any possible (i, i + 3) and (i, i + 4) intrahelicalionic interaction, consists of two identical 35 residue polypeptidechains with a heptad repeat KgLaG-bAcLdEeKf. Three mutant coiled-coilswere prepared where five Glu residues at e positions in EK weremutated to Gin residues (QK); five Lys residues at g positionswere altered to Gin residues (EQ) or these mutations were effectedat both positions e and g (QQ). The stabilities of the fourcoiled-coils were determined by measuring the ellipticitiesat 220 nm as a function of urea concentration at 20C. By usinga double-mutant cycle analysis it was possible to isolate theenergetic contribution of interhelical ionic attractions tocoiled-coil stability from the other contributions such as helicalpreference and hydro-phobicity. The 0.37 0.01 kcal/mol ofenergetic contribution of one interhelical ion pair to the coiled-coilstability was obtained from three independent comparisons. Thisfinding suggests that a large number of weak interhelical electrostaticinteractions on the surface of a protein can make a substantialcontribution to protein stability. In addition, the energeticcontributions of a single mutation E^– Q, K⁺Q, Q E andE^– Ewere also determined (G = 0.22, 0.26, 0.46 and 0.65kcal/mol for the single mutations, respectively). The greatercontribution of a protonated Glu residue to coiled-coil stabilitycompared with an ionized Glu residue (0.65 kcal/mol) can outweighthe relatively smaller contribution of an interhelical ion pair(0.37 kcal/mol), which clearly explains why most coiled-coilsare more stable at acidic pH compared with neutral pH even wheninterhelical salt bridges contribute to the coiled-coil stabilityat neutral pH.     

Deletion of the four C-terminal residues of PepC converts an aminopeptidase into an oligopeptidase

The aminopeptidase PepC is a cysteine peptidase isolated fromlactic acid bacteria. Its structural and enzymatic propertiesclosely resembles those of the bleomycin hydrolases, a groupof cytoplasmic enzymes isolated from eukaryotes. Previous biochemicaland structural data have shown that the C-terminal end of PepCpartially occupies the active site cleft. In this work the substratespecificity of PepC was engineered by deletion of the four C-terminalresidues. The mutant PepC432–435 cleaved peptide substratesas an oligopeptidase while the aminopeptidase specificity wastotally abolished. The substrate size dependency indicated thatPepC432–435 possesses an extended binding site able toaccommodate four residues of the substrate on both sides ofthe cleaved bond. The activity of PepC432–435 towardstryptic fragments of casein revealed a preference for peptideswith hydrophobic amino acids at positions P2 and P3 and forGly, Asn and Gln at position P1. PepC432–435 was shownto be highly sensitive to the thiol peptidase inhibitors leupeptinor E64 which are inefficient towards the wild-type PepC. Inconclusion, deletion of the four C-terminal residues in PepCproduces a new enzyme with properties resembling those of anendopeptidase from the papain family.     

A point mutation of human interferon {gamma} abolishes receptor recognition

We identified a single amino acid mutation that abolished thebioactivity of human IFN. The mutation was identified by screeninga mutagenized IFN expression library for molecules with alteredbiological activity. The mutant protein was expressed at highlevels in Escherichia coli, and remained soluble upon purification.However, the protein was completely inactive in all IFN assaysinvestigated, exhibiting < 0.0006% of the specific activityof native IFN antiviral activity. Sequencing the plasmid DNAencoding this mutant protein showed that the histidine at position111 of native human IFN is changed to aspartic acid (IFN/H111D).Other mutations at this site showed that only hydrophobic aminoacids at position 111 maintain significant, though low, biologicalactivity. Structural characterization of the IFN/H111D proteinby NMR as well as CD spectroscopy demonstrated that the proteinhas limited conformational differences from native IFN. Modelsof the X-ray crystal structure of human IFN [Ealick, P.E., W.J.Cook,S.Vijay-Kumar, M.Carson, T.L.Nagabhushan, P.P.Trotta and C.E.Bugg(1991) Science, 252, 698–702] suggest that this histidineresidue is located at a severe 55° bend in the C-terminalF helix. We conclude that H111 lies within or affects the receptorbinding domain of human IFN.     

Expression of recombinant {alpha}Ains-crystallin and not {alpha}A-crystallin inhibits bacterial growth

A-Crystallin and A^ins-crystallin are derived from the A-crystallingene via alternative splicing. They are identical except forthe presence of a polypeptide, 23 amino acids long, encodedby the ‘insert’ exon. Evolutionary logic would suggestthat the insertion of a 23 amino acid peptide in the middleof A-crystallin, a protein evolving more slowly than eitherhistone H1, cytochrome c or hemoglobin, would lead to appreciablestructural and functional changes. However, based on physico-chemicalstudies, it is presently believed that A-crystallin and A^ins-crystallinare functionally equivalent and that the presence of the ‘insert’peptide in A^Ins-crystallin is inconsequential. We report herethat the independent expression of recombinant A^Ins-crystallin,and not A-crystallin, inhibits growth of the bacterial host.These observations were confirmed in co-expression experiments,wherein both the proteins were expressed in the same cell. Interestingly,growth inhibition is reversible. Importantly, the data demonstratethat it is catalytic amounts and not the gross accumulationof A^Ins-crystalline which causes growth inhibition. Given theprior knowledge that A-crystallin and A^Ins-crystallin differby a peptide of 23 amino acids, these data suggest that the‘insert peptide’ in A^Ins-crystallin imparts propertieson this protein that are different from A-crystallin.     

Folding and stability of the active N-terminal domain of tissue inhibitor of metalloproteinases-1 and -2

The truncated forms of tissue inhibitor of metalloproteinase-1and -2 (TIMP-1 and -2), comprising the N-terminal active domain,are ideal molecules for structural analysis by intrinsic fluorescenceas each contains a single conserved tryptophan residue. In thispaper we describe studies on their conformational stability,unfolding/refolding kinetics and the environment of the uniquetryptophan as judged by its fluorescence properties in the nativestate and exposure to an external quencher, acrylamide. Twoforms of TIMP-2 were studied: TIMP-2 T21 derived from the full-lengthcDNA clone isolated from a mixed-tumour library, and TIMP-2A21 containing the highly conserved V18IRAK22 sequence. In allthree TIMP proteins the tryptophan environments in the nativestate appeared to be similar, but substantial differences wereseen in their conformational stabilities and refolding kinetics.TIMP-1 was approximately twice as stable as TIMP-2 T21 and 1.4-foldmore stable than TIMP-2 A21. This stability difference betweenTIMP-1 and TIMP-2 was shown to be independent of N-linked glycosylation.TTMP-1 and TIMP-2 A21 both showed simple two-state refoldingkinetics, whereas TIMP-2 T21 refolding was more complex andbiphasic in character. These differences between TIMP-2 T21and A21 suggest that residue 21 is a structurally importantsite in the TIMP protein.All three truncated molecules can beconsidered as stable independent folding domains ideally suitedfor further structural analysis     

A point mutation that decreases the thermal stability of human interferon {gamma}

We have identified a mutation of human gamma-interferon (IFN)causing a temperature-sensitive phenotype. We used a randomizedoligonucleotide to mutagenize a synthetic human IFN gene, thenscreened the resulting mutants produced in Escherichia colifor proteins with altered biological activity. One mutant proteinselected for detailed characterization exhibited < 0.3% ofthe specific biological activity of native IFN in an antiviralactivity assay performed at 37°C. However, the protein boundthe human IFN receptor with native efficiency at 4°C. Sequencingthe plasmid DNA encoding this protein snowed that the mutationchanged the lysine residue at amino acid 43 to glutamic acid(IFN/K43E). Site-specific mutagenesis at amino acid 43 showedthat this protein's phenotype resulted from positioning a negativecharge at position 43. Structural characterization of IFN/K43Eusing CD demonstrated that the protein had native conformationat 25°C, but assumed an altered conformation at 37°C.IFN/K43E in this altered conformation bound poorly to the IFNreceptor at 37°C, providing a rationale for the mutant'sdecreased antiviral activity.     

Inhibition of {beta}S-chain dependent polymerization by synergistic complementation of contact site perturbations of {alpha}-chain: application of semisynthetic chimeric {alpha}-chains

Mouse _1–30-horse _31–141 chimeric -chain, a semisyntheticsuper-inhibitory -chain, inhibits ß^S-chain dependent polymerizationbetter than both parent -chains. Although contact site sequencedifferences are absent in the _1–30 region of the chimericchain, the four sequence differences of the region _17-22 couldinduce perturbations of the side chains at ₁₆, ₂₀ and ₂₃, thethree contact sites of the region. A synergistic complementationof such contact site perturbation with that of horse _31–141probably results in the super-inhibitory activity of the chimeric-chain. The inhibitory contact site sequence differences, bythemselves, could also exhibit similar synergistic complementation.Accordingly, the polymerization inhibitory activity of Hb Le-Lamentin(LM) mutation [His20()Gln], a contact site sequence difference,engineered into human–horse chimeric -chain has been investigatedto map such a synergistic complementation. Gln20() has littleeffect on the O₂ affinity of HbS, but in human–horse chimeric-chain it reduces the O₂ affinity slightly. In the chimeric-chain, Gln20() increased sensitivity of the ßßcleft for the DPG influence, reflecting a cross-talk betweenthe ₁ß₁ interface and ßß cleft in this semisyntheticchimeric HbS. In the human -chain frame, the polymerizationinhibitory activity of Gln20() is higher compared with horse_1–30, but lower than mouse _1–30. Gln20() synergisticallycomplements the inhibitory propensity of horse _31–141.However, the inhibitory activity of LM–horse chimeric-chain is still lower than that of mouse–horse chimeric-chain. Therefore, perturbation of multiple contact sites inthe _1–30 region of the mouse–horse chimeric -chainand its linkage with the inhibitory propensity of horse _31–141has been now invoked to explain the super-inhibitory activityof the chimeric -chain. The `linkage-map' of contact sites canserve as a blueprint for designing synergistic complementationof multiple contact sites into -chains as a strategy for generatingsuper-inhibitory antisickling hemoglobins for gene therapy ofsickle cell disease.     

Chemical synthesis of {gamma}-echistatin analogues: elucidation of status of C-terminal (45-49)

Three analogues of -echistatin, des(45–49)--echistatin,des(46–49)-y-echistatin and des(47–49)--echistatin,were synthesized by solid-phase methodology and their biologicalactivities were measured and compared. The results reveal thatwithout the C-terminal (45–49) of -echistatin, the foldingof the protein to the final active structure is not interferedwith and Lys-45 influences the inhibition of platelet aggregation.     

Engineered disulfide bonds in recombinant human interferon-{gamma}: the impact of the N-terminal helix A and the AB-loop on protein stability

Insertion sites for cysteines with optimal stereochemistry forthe formation of unstrained disulfide bridges were identifiedin recombinant human interferon- (rhu-IFN-) by computer modelling.We have engineered two different disulfide cross-linked mutants,containing a pair of symmetry-related disulfide bonds, whichstabilize the N-termini of both monomers of the homodimenc protein.Mutations E7C and S69C allow the formation of an intramonomerdisuffide bond between helices A and D. In contrast, the A17Cand H111C mutations lead to a covalent cross-link between bothmonomers. The AB-loop is linked to helix F. The fluorescenceproperties of native and disulfide cross-linked proteins werestudied as a function of guanidine hydrochloride concentration.Melting temperatures (T_m) were calculated from the decreasein CD ellipticity at 220 nm. The induction of the antiviraleffect was measured using A549 fibroblast cells infected withencephalomyocarditis virus. The ability to induce the expressionof the HLA-DR antigen in Colo 205 cells was determined by fluorescence-activatedcell scanning analysis. The stability of both mutants was stronglyenhanced against temperature- and cosolvent-induced unfolding.The T_m of mutant IFN- E7C/S69C was 15°C. All measured biologicalactivities of this mutant were equal to wild type. In the caseof the other mutant IFN- A17C/H111C, the T_m value was 25°C.This mutation abolishes nearly the entire biological activity(<1%) with no detectable changes of secondary structure inthe CD spectrum. Our results illustrate the importance of theN-terminal helix A and the AB-loop for the unfolding pathwayand thermodynamic stability of rhu-IFN-.     

High-level expression, purification, kinetic characterization and crystallization of protein farnesyltransferase -subunit C-terminal mutants

Protein farnesyltransferase (FPT) is a 97 000 Da heterodimericenzyme that catalyzes post-translational farnesylation of manycellular regulatory proteins including p21 Ras. To facilitatethe construction of site-directed mutants, a novel translationallycoupled, two-cistron Escherichia coli expression system forrat FPT has been developed. This expression system enabled yieldsof >5 mg of purified protein per liter of E.coli cultureto be obtained. The E.coli-derived FPT demonstrated an activitycomparable to that of protein isolated from other sources. Thereported expression system was used to construct three ß-subunitC-terminal truncation mutants, 5, 10 and 14, which were designedto eliminate a lattice interaction between the ß-subunitC-terminus of one molecule and the active site of a symmetry-relatedmolecule. Steady-state kinetic analyses of these mutants showedthat deletion up to 14 residues at the C-terminus did not reducethe value of k_cat; however, K_m values for both peptide and FPPincreased 2–3-fold. A new crystalline form of FPT was obtainedfor the 10 C-terminal mutant grown in the presence of the substrateanalogs acetyl-Cys-Val-Ile-Met-COOH peptide and -hydroxyfarnesylphosphonicacid. The crystals diffract to beyond 2.0 Å resolution.The refined structure clearly shows that both substrate analogsadopt extended conformations within the FPT active site cavity.     

Semi-empirical simulation of Zn/Cd binding site preference in the metal binding domains of mammalian metallothionein

Metallothionein, a two-domain protein, naturally binds sevengram atoms of divalent ions such as Zn and Cd. Four of the metals(Ml, M5, M6 and M7) are found in the -domain and three (M2,M3 and M4) in the ß-domain. Previous studies haveshown that metals in the -domain are more readily exchangeable,and the level of avidity is site specific. By semi-empiricalMNDO modified neglect of diatomic overlap calculations, we foundthe tendency of binding energy for Cd to be M3 > M2 >M4 in the ß-cluster and M5 > M7 > Ml, M6 inthe -cluster. Thus, the replacement of Zn by Cd can be expectedto follow the order M4 M2 M3 in the ß-domain andMS M7 M1 or M6 in the -domain. This is reflected by energydifferences computed with a series of simulated structures derivedfrom either X-ray crystallography or NMR coordinates.     

Family of G protein {alpha} chains: amphipathic analysis and predicted structure of functional domains

The G proteins transduce hormonal and other signals into regulationof enzymes such as adenylyl cyclase and retinal cGMP phosphodiesterase.Each G protein contains an subunit that binds and hydrolyzesguanine nucleotides and interacts with ß subunitsand specific receptor and effector proteins. Amphipathic andsecondary structure analysis of the primary sequences of fivedifferent chains (bovine _s, _t1 and _t2, mouse _i, and rat _o)predicted the secondary structure of a composite chain (_avg).The chains contain four short regions of sequence homologousto regions in the GDP binding domain of bacterial elongationfactor Tu (EF-Tu). Similarities between the predicted secondarystructures of these regions in _avg and the known secondary structureof EF-Tu allowed us to construct a three-dimensional model ofthe GDP binding domain of _avg. Identification of the GDP bindingdomain of _avg defined three additional domains in the compositepolypeptide. The first includes the amino terminal 41 residuesof _avg, with a predicted am phipathic helical structure; thisdomain may control binding of the chains to the ßcomplex. The second domain, containing predicted ßstrands and helices, several of which are strongly amphipathic,probably contains sequences responsible for interaction of chains with effector enzymes. The predicted structure of thethird domain, containing the carhoxy terminal 100 amino acids,is predominantly ß sheet with an amphipathic helixat the carboxy terminus. We propose that this domain is reponsiblefor receptor binding. Our model should help direct further experimentsinto the structure and function of the G protein chain.     

Extrapolation to water of kinetic and equilibrium data for the unfolding of barnase in urea solutions

Assumptions about the dependence of protein unfolding on theconcentration of urea have been examined by an extensive surveyof the equilibrium unfolding of barnase and many of its mutantsmeasured by urea denaturation and differential scanning calorimetry.The free energy of equilibrium unfolding and the activationenergy for the kinetics of unfolding of proteins are generallyassumed to change linearly with [urea]. A slight downward curvatureis detected, however, in plots of highly precise measurementsof logjtu versus [urea] (where k_u is the observed rate constantfor the unfolding of barnase). The data fit the equation logkk_u= logk_u^H₂O* + mk_u^*.[urea] – 0.014[urea]², where mk_u^*is a variable which depends on the mutation. The constant 0.014 was measured directly on four destabilized mutants and wildtype, and was also determined from a global analysis of data from>60 mutants of barnase. Any equivalent deviations from linearityin the equilibrium unfolding are small and in the same region,as determined from measurements on 166 mutants. The free energyof unfolding of barnase, G_U–F, appears significantly largerby 1.6 kcal mol^–1 when measured by calorimetry than whendetermined by urea denaturation. However, the changes in G_U–Fon mutation, G_U–F, determined by calorimetry and by ureadenaturation are identical. We show analytically how, hi general,the curvature in plots of activation or equilibrium energiesagainst [denaturant] should not affect the changes of thesevalues on mutation provided measurements are made over the sameconcentration ranges of denaturant and the curvature is independentof mutation.     

Synthesis and characterization of a recombinant fragment of human {alpha}-fetoprotein with antigenic selectivity versus albumin

A DNA sequence coding for human -fetoprotein amino acid sequence38–119 was synthesized and cloned in a bacterial expressionvector. The -fetoprotein sequence was selected as the leasthomologous to albumin, since the two proteins have an overallamino acid identity of %. A chimeric protein was obtained whichwas purified by preparative electrophoresis and characterizedin its primary structure by fast atom bombardment mass spectrometry.About 70% of the -fetoprotein sequence was physically mappedand found to correspond to the amino acids encoded in the syntheticgene. The use of this recombinant protein allowed the selectionof monoclonal antibodies recognizing both the recombinant fragmentand native -fetoprotein. These antibodies should allow the developmentof an immunoassay for -fetoprotein with absolute selectivityversus albumin. This might result in more sensitive clinicaldeterminations, avoiding the possibility of cross-reactions.     

Structure-function analysis of human IL-1{alpha}: identification of residues required for binding to the human type I IL-1 receptor

Using oligonucleotide-directed mutagenesis, the binding siteon human interleukin-1 (IL-1) for the human type I IL-1 receptor(IL-1R) has been analyzed. Substitution of seven amino acids(Arg12, Ile14, Asp60, Asp61, Ile64, Lys96 and Trp109) resultedin a significant loss of binding to the receptor. Based on crystallographicinformation, the side chains of these residues are clusteredin one region of IL-1 and exposed on the surface of the protein.Five of the residues in the IL-1 binding site align with thebinding residues previously determined in human IL-1ß,demonstrating that the type I IL-1R recognizes homologous regionsin both ligands. Unexpectedly, only three of the aligned residuesare identical between IL-1 and IL-1ß. These observationssuggest that the composition of contact residues in the bindingsite is unique for each ligand–receptor complex in theIL-1 system.     

The predicted secondary structure of the G-type glutamineamidotransferase is compatible with TEM-barrel topology

Glutamine amidotransferase (GAT) subunits or domains catalyzean important partial reaction in many complex biosynthetic reactions.The structure of one member of the F-type GATs is known, butthe structure of the unrelated G-type is still unknown. Becausemany protein sequences are available for anthranilate synthasecomponent II (product of the trpG gene), we have predicted itsaverage secondary structure by a joint prediction method [Niermannand Kirschner (1991a) Protein Engng, 4, 359–370]. Thepredicted eight ß-strands and seven -helices followan 8-fold cyclic repetition of a ß-strand-loop--helix-loopmodule with helix 7 missing. This pattern of secondary structuresuggests that the G-type GAT domain has an 8-fold ß-barreltopology, as found first in triose phosphate isomerase (TIM-barrel).This model is supported by the location of known catalyticallyessential residues in loops between (ß-strands and-helices. Evidence from published sequencing and mutationalstudies on selected members of the GAT superfamily (carbamoylphosphate, imidazoleglycerol phosphate, GMP and CTP synthases)support both the secondary structure prediction and the TIM-barreltopology.     

Independent self-assembly of cadmium-binding {alpha}-fragment of metallothionein in Escherichia coli without participation of {beta}-fragment

We examined the independent self-assembly of the - and ß-fragmentsof human metallothionein (MT) into cadmiumbinding conformationin an Escherichia coli expression system, in addition to wild-typeMT expression. The expressed -fragment formed independentlythe structure of a metal-binding cluster without the aid ofthe ß-fragment. The -fragment and wild-type MT expressedin E.coli were purified and analyzed for their biochemical andspectroscopic properties. The apparent cadmium binding of the-fragment was approximately 12-fold greater than that for thewild-type MT, whereas in other respects the studied biochemicalproperties were similar. In contrast, we were unable to obtainany independently expressed ß-fragment as the cadmium-bindingform in this study. Possible explanations for this phenomenonare discussed.     

Construction of multiple copy of {alpha}-domain gene fragment of human liver metallothionein IA in tandem arrays and its expression in transgenic tobacco plants

Metallothioneins (MT) are low molecular weight, cysteinerich,metal-binding proteins. An MT molecule contains two domainswhich appear to act independently—an -domain, which ischaracterized by cadmium-binding, and a ß-domain,which binds preferentially to copper. Based on this conception,DNA duplex encoding the -domain (106 bp) of human MT-I_A wasconstructed from a chemically-synthesized oligomer by repairsynthesis and enzymatic ligation and cloned into pUC19. Thegenes cloned were sequenced and found to be in the correct orderas designed. Synthetic directional adapters were attached tothe terminals of the -domain gene fragment of human MT-I_A toestablish complete control over fragment orientation duringligation. The use of these directional adapters thereby ensuredthe production of multiple copies of the -domain in tandem arrays.The successive -domains were linked by a peptide linker consistingof 10 residues. A chimeric gene containing 12 cloned tandemlyrepeated copies of the 106 bp -domain DNA was introduced intotobacco cells on a disarmed Ti-plasmid of Agrobacterium tumefaciens.A total of 10 different transgenic tobacco plants were generated,of which two showed root and shoot growth unaffected by up to200 mg/l kanamycin and 100 µM cadmium, whereas root growthof control plants was severely inhibited and leaf chlorosisdeveloped on media containing only 10 µM cadmium