Elongation Factor Tu Research Articles

The structure of transducin G αt: More to view than just Ras

Alfred Wittinghofer Max-Planck-lnstitut ftir Molekulare Physiologie Abteilung Strukturelle Biologie 44139 Dortmund Federal Republic of Germany In the age of molecular biology, structural studies can be done on any protein whose cDNA is available, provided that the protein can be expressed in large enough quanti- ties and can be crystalized. Therefore, it is refreshing that the three-dimensional structure of a protein whose prepa- ration begins in the slaughterhouse with the purchase of hundreds of cow eye retinas has finally, after many years of blood and tears, been solved recently by Noel et al. (1993). As in many other cases, the trick used to obtain crystals of transducin a (G,,) complexed with GTPyS was proteolysis to remove the first 25 amino acid residues from the N-terminus (Mazzoni et al., 1991). Following the bacterial protein synthesis elongation fac- tor Tu (EF-Tu) (Jurnak, 1985; LaCour et al., 1985; Berch- told et al., 1993; Kjeldgaard et al., 1993) and ~21”” (Pai et al., 1989, 1990; Milburn et al., 1990), G,t is the third member of the GTP-binding class of proteins to have its three-dimensional structure determined. The crystal struc- ture of transducin a, however, is the first representative of the family of G, proteins. As expected, its G domain, the structure that binds guanine nucleotides and hydrolyzes GTP, is the same as that of EF-TU and p21”“, making it very likely that any protein with the conserved sequence elements of GTP-binding proteins has this topology (Fig- ure 1). inserted into this conserved G domain (amino acid residues 59-172 of G,,) is a region unique to transducin a. that may hold the key to understanding not only its func- tion as a molecular switch but also give us insights into the function of the other G. family members. Transducin is a heterotrimeric G protein (aPr subunits) that transmits the signal received by the photon receptor rhodopsin to activate the intracellular effector, cGMP phosphodiesterase, by binding to one of the two inhibitory y subunits. The resulting decrease in intracellular cGMP reduces the flow of cations through a cGMP-gated channel and hyperpolarizes the retinal rod cell, creating a signal that is detected at the level of a synaptic connection (Cha- bre and Deterre, 1989). Transducin functions as a switch (Bourne et al., 1990, 1991; Gilman, 1987); the heterotri- merit GDP-bound form is in the off state. Photoactivated rIhodopsin, called metarhodopsin II, binds to the trimer, causes GTP to exchange for GDP, and leads to the separa- tion of the active GTP G,, from the BIr heterodimer. Hydro- lysis of GTP by G,, returns the switch to the off state. Comparison of G. and p2F: Functional Differences What do we learn about G,t and about heterotrimeric G proteins in general, and how does this compare with what we know about other signal switch molecules such as ~21”” (Ras) or the protein synthesis factor EF-Tu? This r’ eview will focus on the comparison of the functional and underlying structural distinctions between two important intracellular signal transducers, Gat and ~21”“.

Conformational differences between complexes of elongation factor Tu studied 19F-NMR spectroscopy.

An analogue of elongation factor Tu (EF-Tu) from Escherichia coli was prepared by biosynthetic incorporation of 3-fluorotyrosine. The 19F-NMR spectra of the binary complexes of this protein with GDP, GTP and elongation factor Ts (EF-Ts) and the ternary complexes EF-Tu.GDP.aurodox and EF-Tu.GDP.EF-Ts were measured. EF-Tu contains ten tyrosine residues and all of the complexes studied gave complex 19F spectra with overlapping resonances. EF-Tu.GDP gave a spectrum in which two signals were markedly different from those shown by the other complexes, the two resonances being shifted downfield by at least 3.4 ppm and 0.9 ppm relative to their shifts in the other complexes. Such large downfield shifts can be explained by second-order electric field shielding effects resulting from these two tyrosine residues being in a sterically constrained environment in EF-Tu.GDP and with the steric restraints being released in all of the other complexes. The X-ray diffraction structure of EF-Tu.GDP shows that Tyr87 in the N-terminal domain (domain I) and Tyr309 in the C-terminal domain (domain III) are both buried within the protein and are close to each other: these residues are in regions of EF-Tu previously implicated in the structural changes between EF-Tu.GDP and EF-Tu.GTP by other workers. If these tyrosine residues correspond to the two downfield resonances of the spectra of EF-Tu.GDP, the results from the 19F-NMR would be consistent with these earlier indications that domain I interacts closely with domain III in EF-Tu.GDP and that the amino acids between Gly83 and Gly100 are an important part of this interaction. For all the other complexes studied, these tyrosines are in a less sterically crowded environment consistent with a weaker interaction between the two domains. The 19F-NMR spectrum of the trypsin-cleaved product of EF-Tu.GDP, from which the X-ray diffraction structural data have been obtained, shows no significant differences from the native protein so that trypsin cleavage causes no large changes in the protein's structure.

Probing the reactivity of the GTP- and GDP-bound conformations of elongation factor Tu in complex with the antibiotic GE2270 A.

The activity of Escherichia coli elongation factor Tu (EF-Tu) depends on its GTP- and GDP-bound conformations. In this work we have studied the influence of the antibiotic GE2270 A, a new EF-Tu-specific inhibitor of protein biosynthesis, on these two conformations with respect to the interaction with the various ligands and stability. One molecule of GE2270 A bound per EF-Tu is sufficient to block poly(U)-directed poly(Phe) incorporation. This drug binds stably to both EF-Tu.GTP and EF-Tu.GDP but only affects the reactivity of the GTP-bound conformation that is no longer able to interact efficiently with aminoacyl-tRNA (aa-tRNA) and ribosomes. Consequently, the protection by EF-Tu.GTP of the aminoacyl-ester bond against nonenzymatic hydrolysis is strongly weakened. The intrinsic EF-Tu GTPase is little affected, but its response to aa-tRNA and ribosomes is impaired. The specificity of GE2270 A to the GTP-bound form of EF-Tu is supported by the course of temperature- and urea-dependent inactivation or tryptic digestion. Furthermore, in its presence elongation factor Ts (EF-Ts) can enhance the dissociation of EF-Tu.GDP but not that of EF-Tu.GTP. Unlike kirromycin, this antibiotic can bind to EF-Tu.EF-Ts without causing dissociation of this complex. Evidence was obtained that the EF-Tu binding site for GE2270 A is also distinct from those for kirromycin, pulvomycin, and aa-tRNA, even though this antibiotic functionally interferes with all three of these ligands. GE2270 A appears to interact with a crucial regulatory region of the EF-Tu molecule that is activated in the GTP-bound state.

Asparagine-135 of elongation factor Tu is a crucial residue for the folding of the guanine nucleotide binding pocket

This work studies the structure-function relationships of Asn 135, a residue situated in the GTP binding pocket of elongation factor Tu (EF-Tu). For this purpose we constructed EF-TuN135D/D138N and assayed its reactivity towards various purine nucleotides. We found that EF-TuN135D/D138N had no functional effect with GTP, ATP, XTP and isoGTP. The lack of a productive interaction with isoGTP shows that the Asn 135 side-chain does not recognize the exocyclic keto group of the guanine base. However, EF-TuN 135D/D 138N, whose native conformation is stabilized by either elongation factor Ts or kirromycin, was able to support the enzymatic binding of aa-tRNA to the ribosome in the absence of any nucleotide, when in complex with the antibiotic. Taken together, these results show that Asn 135 is important for the correct folding of the nucleotide binding site and that EF-Tu·kirromycin can mediate the binding of aa-tRNA to the mRNA-programmed ribosomes independently of the native conformation of this site.

Crystal structure of active elongation factor Tu reveals major domain rearrangements.

The crystal structure of intact elongation factor Tu (EF-Tu) from Thermus thermophilus has been determined and refined at an effective resolution of 1.7 A, with incorporation of data extending to 1.45 A. The effector region, including interaction sites for the ribosome and for transfer RNA, is well defined. Molecular mechanisms are proposed for transduction and amplification of the signal induced by GTP binding as well as for the intrinsic and effector-enhanced GTPase activity of EF-Tu. Comparison of the structure with that of EF-Tu-GDP reveals major mutual rearrangements of the three domains of the molecule.

Purification of chloroplast elongation factor Tu and cDNA analysis in tobacco: the existence of two chloroplast elongation factor Tu species.

We have purified a chloroplast elongation factor Tu (EF-Tu) from tobacco (Nicotiana tabacum) and determined its N-terminal amino acid sequence. Two distinct cDNAs encoding EF-Tu were isolated from a leaf cDNA library of N. sylvestris (the female progenitor of N. tabacum) using an oligonucleotide probe based on the EF-Tu protein sequence. The cDNA sequence and genomic Southern analyses revealed that tobacco chloroplast EF-Tu is encoded by two distinct genes in the nuclear genome of N. sylvestris. We designated the corresponding gene products EF-Tu A and B. The mature polypeptides of EF-Tu A and B are 408 amino acids long and share 95.3% amino acid identity. They show 75-78% amino acid identity with cyanobacterial and chloroplast-encoded EF-Tu species.

Studies on the interaction of elfamycin antibiotics with elongation factor‐Tu by mass spectroscopic techniques

AbstractElfamycin antibiotics have been studied by mass spectrometry, using direct‐liquid‐introduction chemical ionization and fast‐atom bombardment (FAB), but these techniques are not suitable for the identification of the interaction product between antibiotic and elongation factor‐Tu (EF‐Tu). A new mass spectrometric technique, matrix‐assisted laser desorption (MALD), has been employed to clearly identify the complex formed between EF‐Tu and kirromycin. Using MALD operated in the positive‐ion mode, molecular weight values of 43 330 and 44 234 Da were obtained for EF‐Tu and the adduct EF‐Tu/kirromycin, respectively; both are close to the theoretically calculated values of 43 225 and 44 060. These measurements suggest that MALD techniques can be used to obtain useful information on EF‐Tu/elfamicyn antibiotic interaction. Other techniques, such as desorption chemical ionization, FAB and thermospray have been used to obtain information on the fragmen‐tation patterns of some elfamycins (kirrotmycin, aurodox, efrotomycin). In particular, the positive‐ion FAB spectrum of efrotomycin obtained by using a matrix of glycerol and water containing K+ ions, contained the quasimolecular ion [M + K]+ at m/z 1184. Thermospray appeared to be the best technique for structural elucidation studies. In fact, positive‐ion spectra of kirromycin, aurodix and efrotomycin are characterized by significant fragmentation ions, while molecular ions are of very low intensity.

Selection for active E. coli tRNA(Phe) variants from a randomized library using two proteins.

In vitro selection was used to isolate active Escherichia coli tRNA(Phe) variants from randomized libraries. Functional tRNAs were first selected by multiple rounds of binding to Escherichia coli phenylalanyl-tRNA synthetase. These variants were then aminoacylated and selected for affinity to elongation factor-Tu. By randomizing potential recognition nucleotides, the importance of residues U20, G34, A35, A36 and U59, previously identified to be required for specific recognition by E. coli phenylalanyl-tRNA synthetase (FRS), was confirmed. However, the sequences of several active variants imply that the wild-type tertiary interactions G10-C25-U45 and A26-G44 are not required for recognition, as previously suggested. Selection of functional tRNAs from a second library randomized at positions normally involved in conserved tertiary interactions revealed new combinations of nucleotides at these positions, suggesting the presence of novel tertiary interactions. In both libraries, active sequences containing deletions were isolated. Taken together, it is clear that FRS is active with substrates having an unexpectedly broad sequence diversity. Finally, the potency of this method is illustrated by the identification of a second class of variants that was isolated by virtue of the presence of an impurity in the FRS preparation.

Sensitivity of elongation factor Tu (EF-Tu) from different bacterial species to the antibiotics efrotomycin, pulvomycin and MDL 62879

The sensitivity of elongation factor Tu (EF-Tu) from different species of bacteria to the EF-Tu-binding antibiotics efrotomycin, pulvomycin and MDL 62879 was tested by measuring the effect of these antibiotics on cell-free protein synthesis systems. EF-Tu from four different Gram-negative species was sensitive to all three antibiotics. Among Gram-positive bacteria, EF-Tu of Bacillus subtilis, Staphylococcus aureus and Enterococcus faecalis was resistant to efrotomycin and less sensitive to pulvomycin than EF-Tu of Gram-negative bacteria. EF-Tus from streptococci were significantly less sensitive than EF-Tus from Gram-negative bacteria to both efrotomycin and pulvomycin. All of the EF-Tus were sensitive to MDL 62879. The same sensitivity pattern emerged from GDP exchange assays, performed with partially purified EF-Tu from different bacterial species and pure Escherichia coli EF-Ts. These results suggest that the site of action of MDL 62879 is more conserved among bacterial species than those of efrotomycin and pulvomycin. Heterogeneity of EF-Tus from different bacterial species was also reflected in differences in their apparent molecular masses estimated by SDS-PAGE. EF-Tus from the Gram-positive species had higher molecular masses than those from all but one of the Gram-negative species.

TRNA fluorescent labeling at 3' end inducing an aminoacyl-tRNA-like behavior.

A fluorescent tRNA derivative labeled at 3'-O position of the ultimate adenosine residue by reaction, under mild conditions, of tRNA with isatoic anhydride [3,1-benzoxazine-2,4(1H)-dione] was obtained. The labeling selectivity was determined by several criteria: digestion with RNase, followed by HPLC of the digest, produces only one labeled nucleoside, identified as 3'-O-anthraniloyladenosine; the ratio of the absorbance at 260 nm to 332 nm also suggests a 1:1 molar ratio between the nucleic acid and the fluorophore; finally, the incapacity of the labeled tRNA to be charged by the specific aminoacyltransferase further demonstrates the engagement of the 3'-O position. Although the 3'-O-anthraniloyl-labeled tRNA does not seem to be functionally active, as far as the aminoacyl charging activity is concerned, surprisingly we found that it is able to form the ternary complex with elongation factor Tu (EF-Tu) and GTP with an affinity consistently higher than uncharged tRNA. From fluorescence anisotropy measurements the ternary complex dissociation constant was estimated as 73 nM for Escherichia coli and 140 nM for yeast anthraniloyl-tRNA(Phe). These results may be interpreted in terms of the particular structure of the anthraniloyl group that makes the labeled tRNA similar to an aminoacyl-tRNA.

Activated conformations of the ras-gene-encoded p21 protein. 2. Comparison of the computed and high-resolution x-ray crystallographic structures of Gly-12 p21.

The ras-oncogene-encoded p21 protein is a G-protein that has been shown to cause the malignant transformation of normal cells and has been implicated in causing human tumors. p21 is thought to be activated by the binding of GTP in place of GDP to the protein. We have previously constructed the three-dimensional structure of the p21 protein bound to GDP from an available alpha-carbon tracing of this protein using a combination of molecular dynamics and energy minimization (Dykes, et al., J. Biomol. Struct. Dynamics, 9:1025-1044). Until the recent publication of the all-heavy-atom x-ray crystallographic molecular coordinates of p21 residues 1-166 bound to a non-hydrolyzable GTP derivative (GppNp), no all-atom structure of the p21 protein has been available in the Brookhaven National Laboratories Protein Data Bank (PDB). In this communication we compare our computed structure for the p21-GDP complex to this x-ray crystal structure. We find that the two structures agree quite closely with one another, the overall RMS deviation for the backbone being 1.47 A and 2.71 A for all of the atoms. We have identified the regions of the protein that are responsible for the most significant deviations between the two structures, i.e., residues 32-40 and 61-74. We find that the main chain in the 32-40 segment deviates significantly from residue 32 to residue 36 and the side chain phenolic rings of residue 32 differ greatly between the two structures. The 61-74 region is the least-well defined region in the whole protein crystallographically having, by far, the highest temperature factor (B-factor). The backbone and side chain conformations in the 61-74 segment differ markedly, the overall RMS deviation being 3.1 A for the backbone and 5.7 A for all atoms. Both of these regions have been found in x-ray crystallographic studies of p21-GDP and p21-GTP complexes to undergo significant changes in conformation upon the binding of GTP in place of GDP to the protein. We have further compared our computed structure of the p21 protein with the x-ray crystal structure with regard to the conformations of individual segments, in particular, structurally conserved sequences (SCR), i.e., those sequences that have structural and sequence homology to corresponding sequences in the related G-protein, bacterial elongation factor Tu(EF-Tu), and variable loop regions.(ABSTRACT TRUNCATED AT 400 WORDS)

Participation of the overproduced elongation factor Tu from Thermus thermophilus in protein biosynthesis of Escherichia coli

The influence of the overproduced elongation factor Tu (EF-Tu) from Thermus thermophilus on the protein biosynthesis in Escherichia coli was investigated both in vivo and in vitro. A kirromycin-resistant E. coli strain became sensitive to this antibiotic upon the expression of the tufA-gene of T. thermophilus present on a plasmid. In in vitro translation with components of the kirromycin-resistant E. coli strain the poly(Phe) synthesis stopped when minute amounts of the EF-Tu from T. thermophilus were added. Both results indicate the sensitivity of the T. thermophilus EF-Tu to kirromycin and its participation in the polypeptide synthesis of E. coli.

Toward a model for the interaction between elongation factor Tu and the ribosome.

In the elongation cycle of bacterial protein synthesis the interaction between elongation factor-Tu (EF-Tu).guanosine triphosphate (GTP), aminoacyl-transfer RNA (aa-tRNA), and messenger RNA-programmed ribosomes is associated with the hydrolysis of GTP. This interaction determines the selection of the proper aa-tRNA for incorporation into the polypeptide. In the canonical scheme, one molecule of GTP is hydrolyzed in the EF-Tu-dependent binding of aa-tRNA to the ribosome, and a second molecule is hydrolyzed in the elongation factor-G (EF-G)-mediated translocation of the polypeptide from the ribosomal A site to the P site. Substitution of Asp138 with Asn in EF-Tu changed the substrate specificity from GTP to xanthosine triphosphate and demonstrated that the EF-Tu-mediated reactions involved the hydrolysis of two nucleotide triphosphates for each Phe incorporated. This stoichiometry of two is associated with the binding of the correct aa-tRNA to the ribosome.

A single amino acid substitution in elongation factor Tu disrupts interaction between the ternary complex and the ribosome

Elongation factor Tu (EF-Tu).GTP has the primary function of promoting the efficient and correct interaction of aminoacyl-tRNA with the ribosome. Very little is known about the elements in EF-Tu involved in this interaction. We describe a mutant form of EF-Tu, isolated in Salmonella typhimurium, that causes a severe defect in the interaction of the ternary complex with the ribosome. The mutation causes the substitution of Val for Gly-280 in domain II of EF-Tu. The in vivo growth and translation phenotypes of strains harboring this mutation are indistinguishable from those of strains in which the same tuf gene is insertionally inactivated. Viable cells are not obtained when the other tuf gene is inactivated, showing that the mutant EF-Tu alone cannot support cell growth. We have confirmed, by partial protein sequencing, that the mutant EF-Tu is present in the cells. In vitro analysis of the natural mixture of wild-type and mutant EF-Tu allows us to identify the major defect of this mutant. Our data shows that the EF-Tu is homogeneous and competent with respect to guanine nucleotide binding and exchange, stimulation of nucleotide exchange by EF-Ts, and ternary complex formation with aminoacyl-tRNA. However various measures of translational efficiency show a significant reduction, which is associated with a defective interaction between the ribosome and the mutant EF-Tu.GTP.aminoacyl-tRNA complex. In addition, the antibiotic kirromycin, which blocks translation by binding EF-Tu on the ribosome, fails to do so with this mutant EF-Tu, although it does form a complex with EF-Tu. Our results suggest that this region of domain II in EF-Tu has an important function and influences the binding of the ternary complex to the codon-programmed ribosome during protein synthesis. Models involving either a direct or an indirect effect of the mutation are discussed.

Isolation and functional analysis of histidine-tagged elongation factor Tu.

The study of the structure/function relationships of the Escherichia coli elongation factor Tu (EF-Tu) via mutagenesis has been hampered by difficulties encountered in separating the mutated factor from other proteins, in particular native EF-Tu. Here we describe a novel system for the purification of EF-Tu mutant species, based on metal-ion affinity chromatography. To facilitate rapid and efficient purification we designed a recombinant EF-Tu with an additional C-terminal sequence of one serine and six histidine residues. A cell extract containing the His-tagged EF-Tu (EF-TuHis) is applied to a Ni(2+)-nitrilotriacetic acid column. EF-TuHis can be selectively eluted with an imidazole containing buffer, yielding a preparation of more than 95% purity, free of wild-type EF-Tu. In-vitro and in-vivo functional analyses show that EF-TuHis resembles the wild-type EF-Tu, which makes this one-step isolation procedure a promising tool for the study of the interactions of mutant EF-Tu with the various components of the elongation cycle. The new isolation procedure was successfully applied for the purification of a mutant EF-TuHis with a Glu substitution for Lys237, a residue possibly involved in the binding of aminoacyl-tRNA.

The identification of a domain in Escherichia coli elongation factor Tu that interacts with elongation factor Ts.

A method has been developed to search for the elongation factor Tu (EF-Tu) domain(s) that interact with elongation factor Ts (EF-Ts). This method is based on the suppression of Escherichia coli EF-Tu-dominant negative mutation K136E, a mutation that exerts its effect by sequestering EF-Ts. We have identified nine single-amino acid- substituted suppression mutations in the region 146-199 of EF-Tu. These mutations are R154C, P168L, A174V, K176E, D181G, E190K, D196G, S197F, and I199V. All suppression mutations but one (R154C) significantly affect EF-Tu's ability to interact with EF-Ts under equilibrium conditions. Moreover, with the exception of mutation A174V, the GDP affinity of EF-Tu appears to be relatively unaffected by these mutations. These results suggest that the domain of residues 154 to 199 on EF-Tu is involved in interacting with EF-Ts. These suppression mutations are also capable of suppressing dominant negative mutants N135D and N135I to various degrees. This suggests that dominant negative mutants N135D and N135I are likely to have the same molecular basis as the K136E mutation. The method we have developed in this study is versatile and can be readily adapted to map other regions of EF-Tu. A model of EF-Ts-catalyzed guanine-nucleotide exchange is discussed.

The sequence of the gene encoding elongation factor Tu from Chlamydia trachomatis compared with those of other organisms

Nucleotide (nt) sequences encoding the elongation factor Tu(EF-Tu), tRNA The and tRNA Trp from Chlamydia trachomatis have been determined. The environment of the EF-Tu-encoding gene ( tuf), between two tRNA gene sequences, suggests that it is part of a tufB locus. The nt sequence and the deduced amino acid (aa) sequence were aligned with comparable sequences from other organisms and the resulting data bases were used to infer phylogenies. Phylogenetic trees based on aa sequences and nt sequences are similar, but not completely congruent with rRNA gene-based phylogenies. Both the nt and aa sequence trees concur on the early divergence of Thermotoga and Chlamydia from the bacterial root. The aa alignment highlights the presence of four unique Cys residues in the chlamydial sequence which are found at strictly conserved positions in other sequences. Further peculiarities of the chlamydial and eubacterial sequences have been mapped to the X-ray crystallographic structure of the protein.

Sequence of a cDNA encoding the α-subunit of wheat translation elongation factor 1

A cDNA encoding the α-subunit of wheat protein synthesis elongation factor 1 (EF-1α) was isolated from a wheat cDNA expression library and sequenced. The deduced amino acid sequence is compared to EF-1α from other species and to elongation factor Tu (EF-Tu) from Escherichia coli. Putative GTP-binding sites are identified.

Sequence of the tufA gene encoding elongation factor EF-Tu from Thermus aquaticus and overproduction of the protein in Escherichia coli.

The sequence of the tufA gene from the extreme thermophilic eubacterium Thermus aquaticus EP 00276 was determined. The GC content in third positions of codons is 89.5%, with an unusual predominance of guanosine (60.7%). The derived protein sequence differs from tufA- and tufB-encoded sequences for elongation factor Tu (EF-Tu) of Thermus thermophilus HB8, another member of the genus Thermus, in 10 of the 405 amino acid residues. Three exchanges are located in the additional loop of ten amino acids (182-191). The loop, probably involved in nucleotide binding, is absent in EF-Tu of the mesophile Escherichia coli. Since EF-Tu from E. coli is quite unstable, the protein is well-suited for analyzing molecular changes that lead to thermostabilization. Comparison of the EF-Tu domain I from E. coli and Thermus strains revealed clustered amino acid exchanges in the C-terminal part of the first helix and in adjacent residues of the second loop inferred to interact with the ribosome. Most other exchanges in the guanine nucleotide binding domain are located in loops or nearest vicinity of loops suggesting their importance for thermostability. The T. aquaticus EF-Tu was overproduced in E. coli using the tac expression system. Identity of the recombinant T. aquaticus EF-Tu was verified by Western blot analysis, N-terminal sequencing and GDP binding assays.

Optimizing protein crystallization by aggregate size distribution analysis using dynamic light scattering

Dynamic light scattering (DLS) techniques have been used on several systems to distinguish between amorphous and crystalline aggregation of proteins during precipitation. Most of these studies suffer from either of two defects: (i) they have been confined to model protein systems where crystallization conditions are well established, and some cannot be considered as typical proteins in their crystallization behavior, and (ii) the analysis method used to evaluate the DLS results, using a first-cumulant method to give an average hydrodynamic radius, is inappropriate for highly polydisperse systems such as protein solutions containing the free protein together with large aggregates. We present dynamic light scattering (DLS) studies on solutions of aminoacyl-tRNA synthetases as they approach the crystallization point. The systems studied were asparaginyl- (NRSEC), leucyl- (LRSEC), and valyl-tRNA synthetase (VRSEC) from E. coli. As a control system, we used the E. coli elongation factor Tu (EF-Tu). NRSEC, LRSEC, and VRSEC had not been crystallized before. Apart from the different proteins used here, the methods that we have employed differ from previous studies in that (i) instead of making a series of measurements on individual samples at various concentrations, the protein solutions were titrated with the precipitants, and (ii) the results of the light scattering measurements were analyzed by a new maximum entropy procedure that calculates a particle size distribution in a highly reproducible way. The particle size distributions of protein solutions titrated with precipitants are bimodal in most cases. For both peaks, relative areas and mean diffusion coefficients were determined. From comparing the results on the proteins known to crystallize (EF-Tu) with the amorphously precipitating systems (LRS, NRS) we find two necessary, but not sufficient, conditions for the formation of crystals: (i) the diffusion coefficient of the monomer peak stays constant until very close to the precipitation point and (ii) the percentage of large aggregates stays small ( < 10% of the scattered light intensity) during the titration. For VRS both ammonium sulphate and sodium citrate showed a low percentage of large aggregates and a constant diffusion coefficient of the main (protein monomer) peak below the precipitation point. This indicates that both would be possible precipitants for the crystallization of this enzyme. Although no condition could as yet be found for obtaining crystals with ammonium sulphate solutions, crystals of the enzyme have been obtained with sodium citrate. These grow as multicrystals of fine needles; studies are underway to obtain crystals more suitable for diffraction analysis.