The accessibility of antigenic determinants of ribosomal protein s4 in situ

Antibodies to Escherichia coli ribosomal protein S4 react with S4 in subribosomal particles, eg, the complex of 16S RNA with S4, S7, S8, S15, S16, S17, and S19 and the RI* reconstitution intermediate, but they do not react with intact 30S subunits. Antibodies were isolated by three different methods from antisera obtained during the immunization of eight rabbits. Some of these antibody preparations, which contained contaminant antibodies directed against other ribosomal proteins, reacted with subunits, but this reaction was not affected by removal of the anti-S4 antibody population. Other antibody preparations did not react with subunits. It is concluded that the antigenic determinants of S4 are accessible in some protein deficient subribosomal particles but not in intact 30S subunits.     

Binding of erythromycin to the 50S ribosomal subunit is affected by alterations in the 30S ribosomal subunit

Summary Expression of resistance to erythromycin in Escherichia coli, caused by an altered L4 protein in the 50S ribosomal subunit, can be masked when two additional ribosomal mutations affecting the 30S proteins S5 and S12 are introduced into the strain (Saltzman, Brown, and Apirion, 1974). Ribosomes from such strains bind erythromycin to the same extent as ribosomes from erythromycin sensitive parental strains (Apirion and Saltzman, 1974).Among mutants isolated for the reappearance of erythromycin resistance, kasugamycin resistant mutants were found. One such mutant was analysed and found to be due to undermethylation of the rRNA. The ribosomes of this strain do not bind erythromycin, thus there is a complete correlation between phenotype of cells with respect to erythromycin resistance and binding of erythromycin to ribosomes.Furthermore, by separating the ribosomal subunits we showed that 50S ribosomes bind or do not bind erythromycin according to their L4 protein; 50S with normal L4 bind and 50S with altered L4 do not bind erythromycin. However, the 30s ribosomes with altered S5 and S12 can restore binding in resistant 50S ribosomes while the 30S ribosomes in which the rRNA also became undermethylated did not allow erythromycin binding to occur.Thus, evidence for an intimate functional relationship between 30S and 50S ribosomal elements in the function of the ribosome could be demonstrated. These functional interrelationships concerns four ribosomal components, two proteins from the 30S ribosomal subunit, S5, and S12, one protein from the 50S subunit L4, and 16S rRNA.     

Identification of the amino acid residues of proteins S5 and S8 adjacent to each other in the 30 S ribosomal subunit of Escherichia coli.

When Escherichia coli 30 S ribosomal subunits are reacted with protein-protein bifunctional reagents, a number of protein pairs as well as aggregates containing three or more ribosomal proteins are formed. In the present study we have purified one of the protein pairs obtained by reaction of 30 S ribosomal subunits with either radioactive or nonradioactive dimethylsuberimidate. Following molecular weight determination and ammonolysis, the pair was shown to consist of ribosomal proteins S5 and S8. The "native" structure of the complex was surmised from its capacity to be reconstituted into a biologically active 30 S ribosomal subunit. From peptide maps and primary structure determination of various peptides it was demonstrated that the cross-linking bond between ribosomal proteins S5 and S8 involves primarily the residues Lys-93 of protein S8 and the COOH-terminal lysine (Lys-166) of ribosomal protein S5. This result is substantiated by the finding that a mutant carrying an altered S5 lacking the COOH-terminal lysine yields a greatly reduced amount of S5-S8 cross-link. In addition to the points of cross-linking it was found that Lys-30, Lys-68, and Lys-86 of S8 and Lys-5 of S5 react with dimethylsuberimidate, indicating that these residues are available for reaction and suggesting their topographical localization on the ribosomal surface.     

Efficient reconstitution of functional Escherichia coli 30S ribosomal subunits from a complete set of recombinant small subunit ribosomal proteins.

Previous studies have shown that the 30S ribosomal subunit of Escherichia coli can be reconstituted in vitro from individually purified ribosomal proteins and 16S ribosomal RNA, which were isolated from natural 30S subunits. We have developed a 30S subunit reconstitution system that uses only recombinant ribosomal protein components. The genes encoding E. coli ribosomal proteins S2-S21 were cloned, and all twenty of the individual proteins were overexpressed and purified. Reconstitution, following standard procedures, using the complete set of recombinant proteins and purified 16S ribosomal RNA is highly inefficient. Efficient reconstitution of 30S subunits using these components requires sequential addition of proteins, following either the 30S subunit assembly map (Mizushima & Nomura, 1970, Nature 226:1214-1218; Held et al., 1974, J Biol Chem 249:3103-3111) or following the order of protein assembly predicted from in vitro assembly kinetics (Powers et al., 1993, J MoI Biol 232:362-374). In the first procedure, the proteins were divided into three groups, Group I (S4, S7, S8, S15, S17, and S20), Group II (S5, S6, S9, Sll, S12, S13, S16, S18, and S19), and Group III (S2, S3, S10, S14, and S21), which were sequentially added to 16S rRNA with a 20 min incubation at 42 degrees C following the addition of each group. In the second procedure, the proteins were divided into Group I (S4, S6, S11, S15, S16, S17, S18, and S20), Group II (S7, S8, S9, S13, and S19), Group II' (S5 and S12) and Group III (S2, S3, S10, S14, and S21). Similarly efficient reconstitution is observed whether the proteins are grouped according to the assembly map or according to the results of in vitro 30S subunit assembly kinetics. Although reconstitution of 30S subunits using the recombinant proteins is slightly less efficient than reconstitution using a mixture of total proteins isolated from 30S subunits, it is much more efficient than reconstitution using proteins that were individually isolated from ribosomes. Particles reconstituted from the recombinant proteins sediment at 30S in sucrose gradients, bind tRNA in a template-dependent manner, and associate with 50S subunits to form 70S ribosomes that are active in poly(U)-directed polyphenylalanine synthesis. Both the protein composition and the dimethyl sulfate modification pattern of 16S ribosomal RNA are similar for 30S subunits reconstituted with either recombinant proteins or proteins isolated as a mixture from ribosomal subunits as well as for natural 30S subunits.     

Surface topography of the Bacillus stearothermophilus ribosome.

Summary The surface topography of the intact 70S ribosome and free 30S and 50S subunits from Bacillus stearothermophilus strain 2184 was investigated by lactoperoxidase-catalyzed iodination. Two-dimensional polyacrylamide gel electrophoresis was employed to separate ribosomal proteins for analysis of their reactivity. Free 50S subunits incorporated about 18% more ¹²⁵I than did 50S subunits derived from 70S ribosomes, whereas free 30S subunits and 30S subunits derived from 70S ribosomes incorporated similar amounts of ¹²⁵I. Iodinated 70S ribosomes and subunits retained 62–78% of the protein synthesis activity of untreated particles and sedimentation profiles showed no gross conformational changes due to iodination. The proteins most reactive to enzymatic iodination were S4, S7, S10 and Sa of the small subunit and L2, L4, L5/9, L6 and L36 of the large subunit. Proteins S2, S3, S7, S13, Sa, L5/9, L10, L11 and L24/25 were labeled substantially more in the free subunits than in the 70S ribosome. Other proteins, including S5, S9, S12, S15/16, S18 and L36 were more extensively iodinated in the 70S ribosome than in the free subunits. The locations of tyrosine residues in some homologus ribosomal proteins from B. stearothermophilus and E. coli are compared.     

Influence of magnesium and polyamines on the reactivity of individual ribosomal subunit proteins to lactoperoxidase-catalyzed iodination.

30S and 50S subunits, in the presence of either 20 mM Mg2+ or 6 mM Mg2+ and 5mM spermidine plus 25 mM putrescine, were observed to completely associate to form 70S monosomes as monitored by sucrose gradient sedimentation. Subunits maintained under the above ionic conditions were compared with 30S and 50S particles at low (6 mM) magnesium concentration with respect to the reactivity of individual ribosomal proteins to lactoperoxidase-catalyzed iodination. Altered reactivity to enzymatic iodination of ribosomal proteins S4, S9, S10, S14, S17, S19, and S20 in the small subunit of ribosomal proteins, L2, L9, L11, L27, and L30 in the large subunit following incubation with high magnesium or magnesium and polyamines suggests that a conformation change in both subunits accompanies the formation of 70S monosomes. The results further demonstrate that the effect of Mg2+ on subunit conformation is mimicked when polyamines are substituted for magnesium necessary for subunit association.     

Binding of 16S rRNA to chloroplast 30S ribosomal proteins blotted on nitrocellulose

Protein-RNA associations were studied by a method using proteins blotted on a nitrocellulose sheet. This method was assayed with Escherichia Coli 30S ribosomal components. In stringent conditions (300 mM NaCl or 20° C) only 9 E. coli ribosomal proteins strongly bound to the 16S rRNA: S4, S5, S7, S9, S12, S13, S14, S19, S20. 8 of these proteins have been previously found to bind independently to the 16S rRNA. The same method was applied to determine protein-RNA interactions in spinach chloroplast 30S ribosomal subunits. A set of only 7 proteins was bound to chloroplast rRNA in stringent conditions: chloroplast S6, S10, S11, S14, S15, S17 and S22. They also bound to E. coli 16S rRNA. This set includes 4 chloroplast-synthesized proteins: S6, S11, S15 and S22. The core particles obtained after treatment by LiCl of chloroplast 30S ribosomal subunit contained 3 proteins (S6, S10 and S14) which are included in the set of 7 binding proteins. This set of proteins probably play a part in the early steps of the assembly of the chloroplast 30S ribosomal subunit.     

The subunit interface of the Escherichia coli ribosome. Crosslinking of 30 S protein S9 to proteins of the 50 S subunit.

Purified 50 S ribosomal subunits were found to contain significant amounts of protein coincident with the 30 S proteins S9 and/or S11 on two-dimensional polyacrylamide/urea electropherographs. Peptide mapping established that the protein was largely S9 with smaller amounts of S11. Proteins S5 and L6 were nearly coincident on the two-dimensional polyacrylamide/urea electropherographs. Peptide maps of material from the L6 spot obtained from purified 50 S subunits showed the presence of significant amounts of the peptides corresponding to S5. Experiments in which ³⁵S-labelled 30 S subunits and non-radioactive 50 S subunits were reassociated to form 70 S ribosomes showed that some radioactive 30 S protein was transferred to the 50 S subunit. Most of the transferred radioactivity was associated with two proteins, S9 and S5. Sulfhydryl groups were added to the 50 S subunit by amidination with 2-iminothiolane (methyl 4-mercaptobutyrimidate). These were oxidized to form disulfide linkages, some of which crosslinked different proteins of the intact 50 S ribosomal subunit. Protein dimers were partially fractionated by sequential salt extraction and then by electrophoresis of each fraction in polyacrylamide gels containing urea. Slices of the gel were analysed by two-dimensional polyacrylamide/sodium dodecyl sulfate diagonal gel electrophoresis. Final identification of the constituent proteins in each dimer by two-dimensional polyacrylamide/urea gel electrophoresis showed that 50 S proteins L5 and L27 were crosslinked to S9. The evidence suggests that proteins S5, S9, S11, L5 and L27 are located at the interface region of the 70 S ribosome.     

Letter: Conformational changes in ribosomal subunits following dissociation of the Escherichia coli 70 S ribosome

The reactivity of various Escherichia coli ribosomal proteins with N-ethylmaleimide has been used as a probe for ribosomal topography changes during the subunit-70 S transition. With the 70 S ribosome there are several proteins from both subunits which do not react with N-ethylmaleimide, but which do so after dissociation of the 70 S particle to free 30 S and 50 S subunits. The kinetics of their exposure is slow relative to that of the 70 S dissociation reaction suggesting conformational changes in both subunits subsequent to 70 S particle dissociation.     

Characterization of cytoplasmic and chloroplast ribosomal proteins of Euglena gracilis.

Active cytoplasmic ribosone subunits 41 and 62S were prepared by treatment with 0.1 mM puromycin in the presence of 265 mM KCl. Active chloroplast subunits 32 and 49S were obtained after dialysis of chloroplast ribosomal preparations against 1 mM Mg(2+)-containing buffer. Proteins from these different ribosomal particles were mapped by two-dimensional gel electrophoresis in the presence of urea. The 41S small cytoplasmic ribosomal subunit contains 33-36 proteins, the 62S large cytoplasmic ribosomal subunit contains 37-43, the 32S small chloroplast ribosomal subunit contains 22-24, and the 49ts large chloroplast ribosomal subunit contains 30-34 proteins. Since some proteins are lost during dissociation of monosomes into subunits, the 89S cytoplasmic monosome would have 73-83 proteins and the 68S chloroplast monosome, 56-60. The amino acid composition of ribosomal proteins shows differences between chloroplast and cytoplasmic ribosomes.     

Ribosomal protein-dependent orientation of the 16 S rRNA environment of S15

Ribosomal protein S15 binds specifically to the central domain of 16 S ribosomal RNA (16 S rRNA) and directs the assembly of four additional proteins to this domain. The central domain of 16 S rRNA along with these five proteins form the platform of the 30 S subunit. Previously, directed hydroxyl radical probing from Fe(II)-S15 in small ribonucleoprotein complexes was used to study assembly of the central domain of 16 S rRNA. Here, this same approach was used to understand the 16 S rRNA environment of Fe(II)-S15 in 30 S subunits and to determine the ribosomal proteins that are involved in forming the mature S15-16 S rRNA environment. We have identified additional sites of Fe(II)-S15-directed cleavage in 30S subunits compared to the binary complex of Fe(II)-S15/16 S rRNA. Along with novel targets in the central domain, sites within the 5' and 3' minor domains are also cleaved. This suggests that during the course of 30S subunit assembly these elements are positioned in the vicinity of S15. Besides the previously determined role for S8, roles for S5, S6+S18, and S16 in altering the 16 S rRNA environment of S15 were established. These studies reveal that ribosomal proteins can alter the assembly of regions of the 30 S subunit from a considerable distance and influence the overall conformation of this ribonucleoprotein particle.     

Comparative electron microscopic study on the location of ribosomal proteins S3 and S7 on the surface of the E. coli 30S subunit using monoclonal and conventional antibody

Summary Mice were immunised with 30S subunits from E. coli and their spleen cells were fused with myeloma cells. From this fusion two monoclonal antibodies were obtained, one of which was shown to be specific for ribosomal protein S3, the other for ribosomal protein S7. The two monoclonal antibodies formed stable complexes with intact 30S subunits and were therefore used for the three-dimensional localisation of ribosomal proteins S3 and S7 on the surface of the E. coli small subunit by immuno electron microscopy. The antibody binding sites determined with the two monoclonal antibodies were found to lie in the same area as those obtained with conventional antibodies. Both proteins S3 and S7 are located on the head of the 30S subunit, close to the one-third/two-thirds partition. Protein S3 is located just above the small lobe, whereas protein S7 is located on the side of the large lobe.     

Study of the surface of Escherichia coli ribosomes and ribosomal particles by the tritium bombardment method

A new technique of atomic tritium bombardment has been used to study the surface topography of Escherichia coli ribosomes and ribosomal subunits. The technique provides for the labeling of proteins exposed on the surface of ribosomal particles, the extent of protein labeling being proportional to the degree of exposure. The following proteins were considerably tritiated in the 70S ribosomes: S1, S4, S7, S9 and/or S11, S12 and/or L20, S13, S18, S20, S21, L1, L5, L6, L7/L12, L10, L11, L16, L17, L24, L26 and L27. A conclusion is drawn that these proteins are exposed on the ribosome surface to an essentially greater extent than the others. Dissociation of 70S ribosomes into the ribosomal subunits by decreasing Mg2+ concentration does not lead to the exposure of additional ribosomal proteins. This implies that there are no proteins on the contacting surfaces of the subunits. However, if a mixture of subunits has been subjected to centrifugation in a low Mg2+ concentration at high concentrations of a monovalent cation, proteins S3, S5, S7, S14, S18 and L16 are more exposed on the surface of the isolated 30S and 50S subunits than in the subunit mixture or in the 70S ribosomes. The exposure of additional proteins is explained by distortion of the native quaternary structure of ribosomal subunits as a result of the separation procedure. Reassociation of isolated subunits at high Mg2+ concentration results in shielding of proteins S3, S5, S7 and S18 and can be explained by reconstitution of the intact 30S subunit structure.     

Proteins of animal ribosomes. XXV. Proteins of rat liver ribosomes protected by polyuridylic acid from methyl acetimidate substitution.

Addition of poly(U) to complexes of 40S and 60S subunits of rat liver ribosomes decreases the substitution of amino groups of 12 proteins of the small ribosomal subunit and of 11 proteins of the large subunit by [14C]-methyl acetimidate. When comparing the results obtained with this amino group specific reagent with the reactivity of the proteins against iodoacetamide it becomes obvious that 4 proteins of the small ribosomal subunit (S12, 18, 19, 24) and 3 proteins of the large one (L20, 22, 25) are partially protected by poly(U) against reaction with both reagents.     

Protein-RNA crosslinking in Escherichia coli 30S ribosomal subunits. Identification of a 16S rRNA fragment crosslinked to protein S12 by the use of the chemical crosslinking reagent 1-ethyl-3-dimethyl-aminopropylcarbodiimide.

1-ethyl-3-dimethyl aminopropylcarbodiimide (EDC) was used to cross-link 30S ribosomal proteins to 16S rRNA within the E. coli 3OS ribosomal subunit. Covalently linked complexes containing 30S proteins and 16S rRNA, isolated by sedimentation of dissociated crosslinked 30S subunits through SDS containing sucrose gradients, were digested with RNase T1, and the resulting oligonucleotide-protein complexes were fractionated on SDS containing polyacrylamide gels. Eluted complexes containing 30S proteins S9 and S12 linked to oligonucleotides were obtained in pure form. Oligonucleotide 5'terminal labelling was successful in the case of S12 containing but not of the S9 containing complex and led to identification of the S12 bound oligonucleotide as CAACUCG which is located at positions 1316-1322 in the 16S rRNA sequence. Protein S12 is crosslinked to the terminal G of this heptanucleotide.     

Mitochondrial ribosome assembly in Neurospora. Structural analysis of mature and partially assembled ribosomal subunits by equilibrium centrifugation in CsCl gradients

In Neurospora, one protein associated with the mitochondrial small ribosomal subunit (S-5, Mr 52,000) is synthesized intramitochondrially and is assumed to be encoded by mtDNA. When mitochondrial protein synthesis is inhibited, either by chloramphenicol or by mutation, cells accumulate incomplete mitochondrial small subunits (CAP-30S and INC-30S particles) that are deficient in S-5 and several other proteins. To gain additional insight into the role of S-5 in mitochondrial ribosome assembly, the structures of Neurospora mitochondrial ribosomal subunits, CAP-30S particles, and INC-30S particles were analyzed by equilibrium centrifugation in CsCl gradients containing different concentrations of Mg+2. The results show (a) that S-5 is tightly associated with small ribosomal subunits, as judged by the fact that it is among the last proteins to be dissociated in CsCl gradients as the Mg+2 concentration is decreased, and (b) that CAP-30S and INC-30S particles, which are deficient in S-5, contain at most 12 proteins that are bound as tightly as in mature small subunits. The CAP-30S particles isolated from sucrose gradients contain a number of proteins that appear to be loosely bound, as judged by dissociation of these proteins in CsCl gradients under conditions in which they remain associated with mature small subunits. The results suggest that S-5 is required for the stable binding of a subset of small subunit ribosomal proteins.     

Changes in the reactivity of proteins of rat liver ribosomes against [C]iodoacetamide depending on their organization in ribosomal subparticles, 80S ribosomes and on the attachment of poly-(U)

(1) The isolated mixtures of ribosomal proteins can be substituted by [¹⁴C]-iodoacetamide up to an average of about 2 equivalents per 20 000 dalton. The extent of substitution of single proteins measured after two-dimensional polyacrylamide gel electrophoresis shows that all proteins are reactive.

(2) Also in the subunits, all proteins are accessible to substitution. Compared with isolated proteins, however, the reactivity is decreased and the amount of labelling for most proteins ranges as low as 5 to 20%.

(3) Reassociation of ribosomal subunits decreases the reactivity of 12 proteins of the small subunit and that of 20 proteins of the large subunit.

(4) The presence of messenger inhibits the substitution of 10 proteins of the small subunit and of 6 proteins of the large one.

(5) Seven proteins of the small subunit and 3 proteins of the large one are influenced both by the other subunit and by messenger-RNA.     

30S ribosomal subunits can be assembled in vivo without primary binding ribosomal protein S15

Assembly of 30S ribosomal subunits from Escherichia coli has been dissected in detail using an in vitro system. Such studies have allowed characterization of the role for ribosomal protein S15 in the hierarchical assembly of 30S subunits; S15 is a primary binding protein that orchestrates the assembly of ribosomal proteins S6, S11, S18, and S21 with the central domain of 16S ribosomal RNA to form the platform of the 30S subunit. In vitro S15 is the sole primary binding protein in this cascade, performing a critical role during assembly of these four proteins. To investigate the role of S15 in vivo, the essential nature of rpsO, the gene encoding S15, was examined. Surprisingly, E. coli with an in-frame deletion of rpsO are viable, although at 37 degrees C this DeltarpsO strain has an exaggerated doubling time compared to its parental strain. In the absence of S15, the remaining four platform proteins are assembled into ribosomes in vivo, and the overall architecture of the 30S subunits formed in the DeltarpsO strain at 37 degrees C is not altered. Nonetheless, 30S subunits lacking S15 appear to be somewhat defective in subunit association in vivo and in vitro. In addition, this strain is cold sensitive, displaying a marked ribosome biogenesis defect at low temperature, suggesting that under nonideal conditions S15 is critical for assembly. The viability of this strain indicates that in vivo functional populations of 70S ribosomes must form in the absence of S15 and that 30S subunit assembly has a plasicity that has not previously been revealed or characterized.     

Assembly of the 30S subunit from Escherichia coli ribosomes occurs via two assembly domains which are initiated by S4 and S7

A protein which initiates assembly of ribosomes is defined as a protein which binds to the respective rRNA without cooperativity (i.e., without the help of other proteins) during the onset of assembly and is essential for the formation of active ribosomal subunits. The number of proteins binding without cooperativity was determined by monitoring the reconstitution output of active particles at various inputs of 16S rRNA, in the presence of constant amounts of 30S-derived proteins (TP30): This showed that only two of the proteins of the 30S subunit are assembly-initiator proteins. These two proteins are still present on a LiCl core particle comprising 16S rRNA and 12 proteins (including minor proteins). The 12 proteins were isolated, and a series of reconstitution experiments at various levels of rRNA excess demonstrated that S4 and S7 are the initiator proteins. Pulse-chase experiments performed during the early assembly with 14C- and 3H-labeled TP30 and the determination of the 14C/3H ratio of the individual proteins within the assembled particles revealed a bilobal structure of the 30S assembly: A group of six proteins headed by S4 (namely, S4, S20, S16, S15, S6, and S18) resisted the chasing most efficiently (S4 assembly domain). None of the proteins depending on S7 during assembly were found in this group but rather in a second group with intermediate chasing stability [S7 assembly domain; consisting of S7, S9, (S8), S19, and S3]. A number of proteins could be fully chased during the early assembly and therefore represent "late assembly proteins" (S10, S5, S13, S2, S21, S1). These findings fit well with the 30S assembly map.(ABSTRACT TRUNCATED AT 250 WORDS)     

Reverse phase high performance liquid chromatography of Escherichia coli ribosomal proteins: standardization of 70 S, 50 S, and 30 S protein chromatograms

We recently described the use of reverse phase high performance liquid chromatography for the separation of the proteins of the 30 S subunit of Escherichia coli ribosomes (Kerlavage, A. R., Kahan, L., and Cooperman, B. S. (1982) Anal. Biochem. 123, 342-348). In the present studies we report improvements in the technique and its extension to the separation of the proteins of the 50 S subunit and of 70 S ribosomes. Using an octadecasilyl silica column and a trifluoroacetic acid/acetonitrile solvent system, the 21 proteins of the 30 S subunit have been resolved into 17 peaks, the 33 proteins of the 50 S subunit into 22 peaks, and the 53 proteins of the 70 S ribosome into 31 peaks. The proteins present in each peak have been identified by polyacrylamide gel electrophoresis, by comparison with previously standardized chromatograms, and by calibration with authentic samples of purified proteins. All of the known ribosomal proteins have been identified on the chromatograms with the exception of L31 and its variant, L31'. Three protein peaks, not corresponding to known ribosomal proteins, have been observed in preparations from the total protein from 50 S subunits and 70 S ribosomes, but the significance of these peaks is unclear. The reverse phase high performance liquid chromatography technique has the potential for purifying all ribosomal proteins, as demonstrated by the increase in resolution we obtain when a peak isolated under standard gradient conditions and containing several proteins is reapplied to the column and eluted with a shallower gradient. Its utility in preparing proteins for functional studies is demonstrated by a reconstitution of active 30 S particles using 30 S proteins prepared by reverse phase high performance liquid chromatography.