Localization of GAT-1 GABA transporter mRNA in rat striatum: cellular coexpression with GAD67 mRNA, GAD67 immunoreactivity, and parvalbumin mRNA

The cellular localization and neurochemical phenotype of cells expressing the GAT-1 GABA transporter was investigated in the adult rat dorsal striatum using single and dual in situ hybridization and immunocytochemical techniques. Cellular sites of GAT-1, GAD67, and parvalbumin mRNAs were visualized using a combination of radioactive and alkaline phosphatase-labeled oligonucleotides and emulsion autoradiography; GAD67 immunoreactivity was detected using a polyclonal antibody (K2) and 3'3"-diaminobenzidine. Two types of GAT-1-positive striatal cells were detected: (1) those expressing an abundance of GAT-1 mRNA, and (2) those expressing low/undetectable amounts of message. This study focused on the striatal cells expressing an abundance of GAT-1 mRNA; these cells accounted for approximately 3-5% of all striatal neurons and were detected scattered sparsely throughout the striatal complex. Dual in situ hybridization and immunocytochemical studies established that all cells enriched in GAT-1 mRNA also expressed high levels of GAD67 mRNA and were strongly GAD67 immunopositive; the converse was also found to be the case, the two hybridization signals having identical distribution patterns. Further dual in situ hybridization studies established that approximately 60% of these high GAD67/GAT-1 cells expressed parvalbumin mRNA, a marker of one population of striatal interneurons, and had an average cross-sectional area of 152.40 microns 2. The chemical phenotype of the remaining 40% of high GAD67/GAT-1 cells was not determined, although the average cross-sectional area of these cells (102.48 microns 2) was significantly smaller than GAT-1/GAD67/parvalbumin cells; these cells were detected in all striatal regions and are likely to correspond to another population of striatal GABAergic interneuron.

Inactivation of individual Ca(2+)-binding sites in the paired EF-hand sites of parvalbumin reveals asymmetrical metal-binding properties

Previously a rat parvalbumin mutant protein PVF102W was constructed with a reporter Trp at position 102 in the middle of the hydrophobic center [Pauls, T. L., et al. (1993) J. Biol. Chem. 268, 20897-20903]. In the present study three new parvalbumin mutant proteins, derived from PVF102W and containing alterations at positions essential for Ca2+ binding in either one of the two Ca(2+)-binding sites (PV-CD and PV-EF) or in both (PV-CD/-EF), were expressed and purified. With the flow dialysis method it was established that both PV-CD and PV-EF bind 1 Ca2+ with affinity constants KCa of 1.1 x 10(7) and 3.2 x 10(6) M-1, respectively. Mg2+ binding, monitored by equilibrium gel filtration in the absence of Ca2+, showed that both mutants bind 1 Mg2+ with KMg = 8 10(4) for PV-CD and 3 x 10(3) M-1 for PV-EF. Compared to the parameters of the parent mutant PVF102W (two sites with equal affinities of 2.7 x 10(7) and 3 x 10(4) M-1 for Ca2+ and Mg2+, respectively), these data indicate that inactivation of the EF site, much more than of the CD site, impairs divalent cation binding. The binding of Ca2+ and Mg2+ is mutually exclusive, indicative of a Ca2+/Mg2+ mixed site. However, as for PVF102W, the KMg values obtained from the competition equation are approximately 40-fold lower than the affinities measured by direct binding. PV-CD/-EF binds neither Ca2+ nor Mg2+. Trp fluorimetry revealed that in the three mutant PVs the residue Trp-102 is deeply buried in the hydrophobic core.(ABSTRACT TRUNCATED AT 250 WORDS)

Cloning of the Saccharomyces cerevisiae gene whose overexpression overcomes the effects of HM-1 killer toxin, which inhibits beta-glucan synthesis

A gene whose overexpression can endow Saccharomyces cerevisiae cells with resistance to HM-1 killer toxin was cloned from an S. cerevisiae genomic library. This gene, designated HKR1 (Hansenula mrakii killer toxin-resistant gene 1), contains a 5.4-kb open reading frame. The predicted amino acid sequence of the protein specified by HKR1 indicates that the protein consists of 1,802 amino acids and is very rich in serine and threonine, which could serve as O-glycosylation sites. The protein also contains two hydrophobic domains at the N-terminal end and in the C-terminal half, which could function as a signal peptide and transmembrane domain, respectively. Hkr1p is found to contain an EF hand motif of the calcium-binding consensus sequence in the C-terminal cytoplasmic domain. Thus, Hkr1p is expected to be a calcium-binding, glycosylated type I membrane protein. Southern and Northern (RNA) analyses demonstrated that there is a single copy of the HKR1 gene in the S. cerevisiae genome, and the transcriptional level of HKR1 is extremely low. Gene disruption followed by tetrad analysis showed that HKR1 is an essential gene. Overexpression of the truncated HKR1 encoding the C-terminal half of Hkr1p made the cells more resistant to HM-1 killer toxin than the full-length HKR1 did, demonstrating that the C-terminal half of Hkr1p is essential for overcoming the effect of HM-1 killer toxin. Furthermore, overexpression of HKR1 increased the beta-glucan content in the cell wall without affecting in vitro beta-glucan synthase activity, suggesting that HKR1 regulates beta-glucan synthesis in vivo.

Characterization of parvalbumin cDNA clones and gene expression in normal and dystrophic mice of strain 129 ReJ

Parvalbumin is a calcium-binding protein found in fast-twitch skeletal muscles and selected cells in the brain. In several dystrophic mutants in the mouse, the parvalbumin content of skeletal muscles and brain is reduced and this deficiency appears to correlate with the inability of these mice to handle enhanced calcium uptake associated with the dystrophic process. In this study, two overlapping cDNA clones of 392 and 1268 base pairs were isolated from a mouse cDNA library in lambda gt11, characterized, and used as probes to study the involvement of the parvalbumin gene and its expression in various tissues of dystrophic mice of strain 129 ReJ. Southern blot analyses of restriction fragments of genomic DNA from normal and dystrophic mice indicate the same number and size of parvalbumin-specific gene fragments observed by other researchers, suggesting that the size of the Pva gene is the same in both normal and dystrophic mice of strain 129 ReJ. Northern blot analyses of total RNA from hind-limb muscles using cloned parvalbumin cDNA as probes revealed an abundant 800-nucleotide mRNA with lesser amounts of a 1000-nucleotide mRNA transcript in both normal and dystrophic mice of strain 129 ReJ. The amount of these mRNAs was reduced by 65-77% in dystrophic muscles preparations and was similar to the levels of beta-actin mRNA in these animals. These results suggest that parvalbumin gene expression is not down regulated in dystrophic mice of strain 129 ReJ.

A new approach to the design of stable proteins

We propose a simple algorithm to design a sequence which fits a given protein structure with a given energy. The algorithm is a modification of the Metropolis Monte Carlo scheme in sequence space with an evolutionary temperature which sets the energy scale. There is a one to one correspondence between this optimization scheme and the Ising model of ferromagnetism. This analogy implies that the design algorithm does not encounter multiple-minima problems and is very fast. The algorithm is tested by 'predicting' the primary structures of four proteins. In each case the calculated primary structures had statistically significant homology with the natural structures.

Metal binding properties of recombinant rat parvalbumin wild-type and F102W mutant

Rat parvalbumin (PV), an EF-hand type Ca(2+)-binding protein, was expressed in Escherichia coli and mutated by replacing a Phe at position 102 with a unique Trp in order to introduce a distinct fluorescent label into the protein. Mass spectroscopy and NMR data indicate that the recombinant wild-type (PVWT) and F102W mutant (PVF102W) proteins have the expected molecular weight and retain the native structure. Both proteins contain two non-cooperative Ca2+/Mg(2+)-binding sites with intrinsic affinity constants, KCa and KMg, of 2.4 +/- 0.9 x 10(7) M-1 and of 2.9 +/- 0.2 x 10(4) M-1, respectively, for PVWT, and KCa and KMg, of 2.7 +/- 1.1 x 10(7) M-1 and of 4.4 +/- 0.3 x 10(4) M-1, respectively, for PVF102W. Based on the highly similar metal binding properties of PVWT and PVF102W the latter protein was used to study cation-dependent conformational changes. Trp fluorescence emission and UV difference spectra of PVF102W indicated that the Trp residue at position 102 is confined to a hydrophobic core and conformationally strongly restricted. Upon Ca2+ or Mg2+ binding the structural organization of the region around the Trp is hardly affected, but there are significant changes in its electrostatic environment. The conformational change upon binding of Ca2+ and Mg2+, as monitored by UV difference spectrophotometry, increases linearly from 0 to 2 cations bound, indicating that the binding of both ions contributes equally to the structural organization in this protein.

Human alpha and beta parvalbumins. Structure and tissue-specific expression

alpha and beta parvalbumins are Ca(2+)-binding proteins of the EF-hand type. We determined the protein sequence of human brain alpha parvalbumin by mass spectrometry and cloned human beta parvalbumin (or oncomodulin) from genomic DNA and preterm placental cDNA. beta parvalbumin differs in 54 positions from alpha parvalbumin and lacks the C-terminal amino acid 109. From MS analyses of alpha and beta parvalbumins we conclude that parvalbumins generally lack posttranslational modifications. alpha and beta parvalbumins were differently expressed in human tissues when analyzed by immunoblotting and polymerase-chain-reaction techniques. Whereas alpha parvalbumin was found in a number of adult human tissues, beta parvalbumin was restricted to preterm placenta. The pattern of alpha parvalbumin expression also differs in man compared to other vertebrates. For example, in rat, alpha parvalbumin was found in extrafusal and intrafusal skeletal-muscle fibres whereas, in man, alpha parvalbumin was restricted to the muscle spindles. Different functions for alpha and beta parvalbumins are discussed.

Specific increase of genetic expression of parvalbumin in fast skeletal muscles of mdx mice

Parvalbumin mRNA was assayed by Northern blot analysis in muscles from normal and dystropic (mdx) mice. Its content was found to be specifically higher in mdx fast muscles than in control preparations. This suggests an increased expression of the protein in dystrophin-lacking fast fibres. A possible role in calcium homeostasis is discussed.

Parvalbumin immunoreactivity in the hippocampus of the gerbil after transient forebrain ischaemia: a qualitative and quantitative sequential study

Parvalbumin immunoreactivity is examined in the hippocampus of the Mongolian gerbil (Meriones unguiculatus) in controls and in animals subjected to 20 min of forebrain ischaemia produced by bilateral clipping of the carotids. In comparison with other species, the hippocampus of the gerbil is characterized by strong immunoreactivity of the (presumably excitatory) perforant pathway, and weak immunoreactivity (low numbers of neurons and scarce dendritic arbors) in nonpyramidal nerve cells (inhibitory neurons) of the CA1 area. These properties may play some role in the development and maintenance of seizures in this susceptible species. Parvalbumin immunoreactivity is rapidly and ephemerally increased in the hippocampus 15 min after reperfusion. Later on, there is a transitory decrease of parvalbumin immunoreactivity which is followed by an increase 6 h later in the stratum granulare hilus and CA3 area, and not until the first and second days in the CA1 area. This increase significantly surpasses the number of immunoreactive neurons in control animals in CA1 and CA3 from 48 h after reperfusion onwards. The effect is similar using different anaesthetics and does not occur in sham-operated animals. In contrast with these findings, the number of parvalbumin-immunoreactive neurons in the somatosensory cortex is not affected in our model of forebrain ischaemia. On the other hand, GABA-immunoreactive neurons in CA1 are preserved during the first week after reperfusion, although an increase in the number of these cells occurs at the end of this period. Delayed neuronal death occurs in the CA1 area 48 h after ischaemia, and marked reduction in the number of CA1 neurons is found by the end of the first week. Eighty per cent of the remaining cells in CA1 at day 7, and 83% at day 15, are parvalbumin-immunoreactive nonpyramidal neurons in contrast to 3% parvalbumin-immunoreactive cells in control animals. These findings indicate that GABAergic neurons in CA1 are preserved after forebrain ischaemia, and that parvalbumin in CA1 neurons is associated with survival.

Evolution of EF-hand calcium-modulated proteins. V. The genes encoding EF-hand proteins are not clustered in mammalian genomes

The chromosomal assignments of genes belonging to the EF-hand family which have a common origin are compiled in this article. So far data are available from 27 human gene loci belonging to 6 subfamilies and 8 murine loci belonging to 4 subfamilies. Chromosomal localization has been obtained by somatic-cell hybrid analysis using the Southern blot technique or PCR amplification, metaphase spread in situ hybridization, or isolation of the particular genes from chromosome-specific libraries. Except for genes of the S-100 alpha proteins which are grouped on human chromosome 1q12-25 and mouse chromosome 3, no linkage has been found for genes encoding EF-hand proteins, indicating absence of selective pressure for maintaining chromosomal clustering. Six of these genes map to known syntenic groups conserved in the human and mouse genomes. This suggests that chromosomal translocations occurred before divergence of these species. The possible significance of chromosomal positioning with respect to nearby located known genes and genetic disease loci is discussed.

Evidence that deprotonation of serine-55 is responsible for the pH-dependence of the parvalbumin Eu3+ 7F0-->5D0 spectrum

The Eu(III)7F0-->5D0 excitation spectra of the parvalbumins are highly pH-dependent. Below pH 6.0, they exhibit a sharp, partially resolved doublet centered near 5,795 A. However, as the pH is raised, the spectrum becomes increasingly dominated by a much broader signal near 5,784 A. This behavior has been traced to the Eu(III) ion bound at the CD site, but the identity of the moiety undergoing deprotonation remains uncertain. Site-specific mutagenesis studies on the parvalbumin-like protein known as oncomodulin now suggest that the species in question is a liganding serine hydroxyl group. Specifically, replacement of serine-55 by aspartate (the residue present at the corresponding position in the EF site) affords a protein that retains two functional lanthanide binding sites, but fails to undergo the pH-dependent spectral alteration. By contrast, replacement of aspartate-59 by glycine (the corresponding EF site residue) fails to abolish the pH-dependent behavior.

Comparison of terbium (III) luminescence enhancement in mutants of EF hand calcium binding proteins

The luminescent isomorphous Ca2+ analogue, Tb3+, can be bound in the 12-amino acid metal binding sites of proteins of the EF hand family, and its luminescence can be enhanced by energy transfer from a nearby aromatic amino acid. Tb3+ can be used as a sensitive luminescent probe of the structure and function of these proteins. The effect of changing the molecular environment around Tb3+ on its luminescence was studied using native Cod III parvalbumin and site-directed mutants of both oncomodulin and calmodulin. Titrations of these proteins showed stoichiometries of fill corresponding to the number of Ca2+ binding loops present. Tryptophan in binding loop position 7 best enhanced Tb3+ luminescence in the oncomodulin mutant Y57W, as well as VU-9 (F99W) and VU-32 (T26W) calmodulin. Excitation spectra of Y57F, F102W, Y65W oncomodulin, and Cod III parvalbumin revealed that the principal Tb3+ luminescence donor residues were phenylalanine or tyrosine located in position 7 of a loop, despite the presence of other nearby donors, including tryptophan. Spectra also revealed conformational differences between the Ca2+- and Tb(3+)-bound forms. An alternate binding loop, based on Tb3+ binding to model peptides, was inserted into the CD loop of oncomodulin by cassette mutagenesis. The order of fill of Tb3+ in this protein reversed, with the mutated loop binding Tb3+ first. This indicates a much higher affinity for the consensus-based mutant loop. The mutant loop inserted into oncomodulin had 32 times more Tb3+ luminescence than the identical synthetic peptide, despite having the same donor tryptophan and metal binding ligands. In this paper, a ranking of sensitivity of luminescence of bound Tb3+ is made among this subset of calcium binding proteins. This ranking is interpreted in light of the structural differences affecting Tb3+ luminescence enhancement intensity. The mechanism of energy transfer from an aromatic amino acid to Tb3+ is consistent with a short-range process involving the donor triplet state as described by Dexter (Dexter, D. L. (1953) J. Chem. Phys. 21, 836). This cautions against the use of the Förster equation in approximating distances in these systems.

Homology modelling of integrin EF-hands. Evidence for widespread use of a conserved cation-binding site

Integrin alpha-subunits contain three or four peptide sequences that are similar to the EF-hand, a 13-residue bivalent cation-binding motif found in calmodulin and parvalbumin. The integrin sequences differ from classical EF-hands in that they lack a co-ordinating residue at position 12. One hypothesis to explain integrin-ligand binding is that aspartate-containing recognition sequences in integrin ligands, which bind at or near to the EF-hand-like sequences, may take the place of the missing residue and co-ordinate directly to the bound cation. In this report, homology modelling of integrin EF-hand-like sequences has been performed using the X-ray structure of calmodulin as a template in order to assess the functional activity of the integrin sequences. In the calmodulin-integrin hybrid structures, integrin EF-hand-like sequences were able to retain cations whereas control sequences did not. Structural analyses demonstrated that the integrin sequences in the hybrid proteins closely resembled conventional EF-hands. The integrin sequences are therefore highly likely to bind Ca2+ ions in vivo, a prerequisite for the ligand-binding model. Database searching with a matrix derived from known integrin EF-hand-like sequences has been used to identify other proteins containing the integrin EF-hand-like motif. Annexin V (anchorin CII), atrial natriuretic peptide receptors and the 70 kDa heat-shock protein were identified by the matrix; the functions of these proteins are known from previous studies to be bivalent cation-dependent. These findings suggest that the integrin EF-hand-like sequence may be a more common motif than originally thought.

An STS in the human parvalbumin gene (PVALB)

Images

The genes for the highly homologous Ca(2+)-binding proteins oncomodulin and parvalbumin are not linked in the human genome

The chromosomal loci of the human parvalbumin and oncomodulin single-copy genes that encode structurally and evolutionarily closely related Ca(2+)-binding proteins were determined by somatic cell hybrid analysis. Southern blot analysis of genomic DNA from 25 human-hamster somatic cell hybrids showed that the human gene for oncomodulin resides on chromosome 7. Analysis of human-mouse hybrids selectively retaining human chromosome 7 or a portion of it allowed specific assignment of the gene locus to the p11-p13 region of chromosome 7 known to be mutated or deleted in patients with the Greig cephalopolysyndactyly syndrome. By gene dosage analysis on Southern blots, we showed that the gene for human parvalbumin maps distally to the cat eye syndrome marker D22S9 on chromosome 22q. Using somatic cell hybrids containing parts of human chromosome 22, the parvalbumin gene was sublocalized to the region 22q12-q13.1. This region contains a linkage group that maps to mouse chromosome 15, region E, and includes the SIS, ARSA, and DIA 1 genes. Our findings are consistent with the recent localization of the mouse parvalbumin gene to this region by two independent methods (C. H. Zühlke et al., 1989, Genet. Res. 54:37-43; S. Adolph et al., 1989, Cytogenet. Cell Genet. 52:177-179).

The structure of the mouse parvalbumin gene

Parvalbumin (PV) is a calcium-binding protein of the EF-hand family, expressed mainly in fast contracting/relaxing muscles of vertebrates. We have isolated five overlapping genomic PV clones which overall span 28 kilobase pairs (kb) around the Pva locus on mouse Chromosome (Chr) 15. The positions of four introns were determined by DNA sequencing. They interrupt the coding sequences at positions corresponding to those in rat and human PV genes. The transcription start site, 25 bp downstream from the TATA-box, was mapped by oligonucleotide primer extension on poly(A)(+)-RNA. The analysis of 0.4 kb promoter sequence of the mouse PV gene revealed CCAAT- and TATA-box sequences and a 59 bp GC-rich stretch between positions -59 and -118. Similar motifs have been found in the parvalbumin genes of rat and human. A perfect 11-bp repeat upstream to positions -149 and -163 respectively is homologous only to the rat promoter. These results will be related to tissue and species differences in PV expression.

Comparison of the amino acid sequences of tissue-specific parvalbumins from chicken muscle and thymus and possible evolutionary significance

Chicken leg muscle parvalbumin was digested with cyanogen bromide or trypsin or trypsin after citraconylation. Peptides isolated by reverse phase HPLC at pH 7.0 were subjected to acid hydrolysis and amino acid analysis and, in some cases, sequencing. The chicken muscle parvalbumin amino acid sequence has ca. 80% sequence identity with alpha-type parvalbumins from mammalian (rabbit, human and rat) muscle. By contrast, the chicken thymus parvalbumin ("avian thymic hormone") sequence is very similar to reptile (turtle, salamander and frog) muscle beta-type parvalbumins. We hypothesize that the evolutionary appearance of the warm-blooded reptiles was accompanied by recruitment of the beta parvalbumin isozyme for promotion of lymphocyte maturation.

Neural regulation of calmodulin in adult Xenopus leg muscle

Two Ca(2+)-binding proteins important in regulating muscle responses to Ca2+ flux are differentially expressed following denervation of Xenopus laevis gastrocnemius. Levels of parvalbumin (PV) RNA transcripts and proteins decrease in abundance, while calmodulin (CaM) transcript and protein levels increase. Our studies on PV kinetics in Xenopus follow a pattern observed in other species, however, our observation of a concomitant increase in CaM has not been documented in any system. Molecular analyses of the Xenopus CaM gene indicate that its structure and upstream sequences are highly conserved across several vertebrate species and implicate several transcription factors in the regulation of its expression.

Interactions between residues in the oncomodulin CD domain influence Ca2+ ion-binding affinity

Despite striking sequence homology with rat parvalbumin, oncomodulin exhibits much lower affinity for Ca2+ ion. We are attempting to identify the structural basis for this difference by systematically substituting the parvalbumin residue for the oncomodulin residue at points of nonidentity. In this paper, we examine two mutations in the helical segments flanking the CD ion-binding loop. Replacement of Asp-45 in the C helix by lysine, to produce D45K, reduces the dissociation constant for Ca2+ at the CD site from 0.81 to 0.53 microM. Replacement of Lys-69 in the D helix by glycine, to afford K69G, similarly reduces KCa to 0.59 microM. Both mutations perturb the Eu3+ 7Fo----5Do spectral parameters. We also examine the consequences of simultaneous mutations involving positions 57, 59, 60, and 69. Ca(2+)-binding assays and Eu3+ luminescence measurements indicate that there is a conformational interaction between residues 57 and 69 and that this interaction is modulated by residues 59 and 60. When the mutations at positions 57, 59, 60, and 69 are combined, the resulting variant exhibits a KCa value for the CD site of 0.25 microM, reflecting a 3-fold increase in affinity relative to the wild-type protein. Moreover, the pK alpha governing the interconversion of low and high pH forms of the Eu3+ 7Fo----5Do spectrum is increased to 8.1, very close to the value of 8.25 determined previously for rat parvalbumin. In this paper, we also complete our survey of single mutations in the CD loop by examining L58I. Replacement of Leu-58 by isoleucine reduces the affinity of the CD site for Ca2+, raising KCa to 2.2 microM. Finally, we revise our previous estimate of the KCa value for Y57F downward, from 0.80 to 0.64 microM. The earlier result is believed to have been inflated by heterogeneity in the preparation, a consequence of proteolysis.

Avian thymic hormone and chicken (muscle) parvalbumin are distinct proteins: isolation of a muscle parvalbumin cDNA fragment by PCR

Access to the nucleotide sequence of parvalbumin from chicken muscle was gained via the polymerase chain reaction. In the absence of specific amino acid sequence data, the PCR primers were based on consensus data for the two parvalbumin Ca2(+)-binding sites. The 137 bp fragment obtained by amplification clearly codes for a parvalbumin, as judged by the presence of 10 invariant codons within the sequence flanked by the primers. When used to probe poly(A)+ RNA from chicken muscle, the fragment recognizes an 800 nucleotide transcript. The translated nucleotide sequence of the muscle protein is unmistakably distinct from that of the thymus-specific parvalbumin known as avian thymic hormone. Of the 31 amplified residues, the two proteins differ at 14. The presence of a distinct parvalbumin in chicken thymus is consistent with the potent effector role proposed for the protein.

Molecular cloning of the thymus-specific parvalbumin known as avian thymic hormone: isolation of a full length cDNA and expression of the recombinant protein in Escherichia coli

The complete coding sequence of the thymus-specific parvalbumin called avian thymic hormone (ATH) has been cloned into Escherichia coli. The translated amino acid sequence was found to be identical to the sequence of map turtle parvalbumin at 90 of 108 positions. Northern blot analysis of thymic RNA indicated a transcript length of approximately 1050 bp. However, the ATH cDNA probe failed to hybridize to poly(A)+ RNA from chicken leg muscle, a further indication that avian thymic hormone is distinct from the muscle-associated parvalbumin previously isolated from chicken. Southern analysis of chicken genomic DNA suggests the presence of a single copy of the ATH gene, and the absence of hybridization between an ATH cDNA fragment and genomic DNA from rat and rabbit is confirmatory evidence that ATH expression is restricted to avian species. One of the full length ATH cDNA clones harbored an insert that lacked all 5' noncoding sequences. This cDNA was inserted without further alteration into the prokaryotic expression vector, pKK223-3. The resulting construction, which contains eleven base pairs between the Shine-Dalgarno sequence and the initiation codon, affords reasonably high levels of expression in E. coli. In most respects, recombinant ATH mimics the tissue-derived protein, retaining a similarly high affinity for Ca2+ ion (KCa = 14 +/- 5 nM). However, in contrast to ATH isolated from chicken thymus tissue, the N-terminal alanine of recombinant ATH is unacetylated. As a result, the isoelectric point is shifted upward from 4.3 to approximately 4.8.

Structure of oncomodulin refined at 1.85 A resolution. An example of extensive molecular aggregation via Ca2+

The crystal structure of oncomodulin, a 12,000 Mr protein isolated from rat tumours, has been determined by molecular replacement using the carp parvalbumin structure as a starting model. Refinement was performed by cycles of molecular fitting and restrained least-squares, using area-detector intensity data to 1.85 A resolution. For the 5770 reflections in the range 6.0 to 1.85 A, which were used in the refinement, the crystallographic R-factor is 0.166. The refined model includes residues 2 to 108, three Ca2+ and 87 water molecules per oncomodulin molecule. The oncomodulin backbone is closely related to that of parvalbumin; however, some differences are found after a least-squares fit of the two backbones, with root-mean-square (r.m.s.) deviations of 1 to 2 A in residues 2 to 6, 59 to 61 of the CD loop, 87, 90 and 108. The overall r.m.s. deviation of the backbone residues 5 to 108 is 0.62 A. Each of the two Ca2+ atoms that are bound to the CD and EF loops is co-ordinated to seven oxygen atoms, including one water molecule. The third Ca2+ is also seven-co-ordinated, to five oxygen atoms belonging to three different oncomodulin molecules and to two water molecules which form hydrogen bonds to a fourth oncomodulin; thus, this intermolecular Ca2+ and its equivalents interlink the molecules into zigzag layers normal to the b axis with a spacing of b/2 or 32.14 A. No such extensive molecular aggregation has been reported for any of the related Ca-binding regulatory proteins of the troponin-C family studied thus far. The Ca-O distances in all three polyhedra are in the range 2.07 A to 2.64 A, indicating tightly bound Ca polyhedra.

Partial nucleotide sequence of the parvalbumin from chicken thymus designated "avian thymic hormone"

The incomplete amino acid sequence of the protein identified as avian thymic hormone was recently reported [Brewer et al. (1989), Biochem. Biophys. Res. Commun. 160, 11555-1161], and a very high degree of homology to the parvalbumins was apparent. Using mixed oligonucleotide primers based on the reported protein sequence, we have succeeded in amplifying and cloning a 188 bp fragment of the coding region for this protein, beginning with double-stranded cDNA prepared from chicken thymus mRNA. The translated nucleotide sequence of this fragment and the reported amino acid sequence display substantial disagreement. Most notably, the nucleotide sequence indicates that the CD site of avian thymic hormone is a typical parvalbumin CD site.

Heat capacity and entropy changes of the major isotype of the toad (Bufo) parvalbumin induced by calcium binding

The possible structural changes in the major isotype of parvalbumin from the toad (Bufo bufo japonicus) skeletal muscle caused by Ca2+ and Mg2+ binding have been analyzed by microcalorimetric titrations. Parvalbumin was titrated with Ca2+ in both the absence and presence of Mg2+ and with Mg2+ in the absence of Ca2+, at pH 7.0, and at 5 degrees, 15 degrees, and 25 degrees C. The two sites in a molecule were equivalent on Mg2(+)-Ca2+ exchange, but distinguishable on Ca2+ and Mg2+ binding. The reactions of parvalbumin with Ca2+ are exothermic at every temperature in both the absence and presence of Mg2+, but those with Mg2+ are always endothermic except for the binding to site 1 at 25 degrees C. The magnitudes of the hydrophobic and internal vibrational contributions to the heat capacity and entropy changes of parvalbumin on Ca2+ and Mg2+ binding and Mg2(+)-Ca2+ exchange have been estimated by the empirical method of Sturtevant [Sturtevant, J. M. (1977) Proc. Natl Acad. Sci. USA 74, 2236-2240]. Although no major conformational changes were noted between Ca2(+)- and Mg2(+)-bound forms of toad parvalbumin, the conformational difference was larger in Ca2+ (or Mg2+) binding to site 1 than site 2. This may indicate that the metal-free form is much less stable than any form with Ca2+ (or Mg2+) bound at one site at least. On Mg2(+)-Ca2+ exchange, the vibrational as well as hydrophobic entropy is only slightly increased in a parallel manner. In contrast, on Ca2+ (or Mg2+) binding, the hydrophobic entropy increases but the vibrational entropy decreases; the former indicates the sequestering of nonpolar groups from the surface to the interior of a molecule, and the latter suggests that the overall structures are tightened on Ca2+ (or Mg2+) binding but loosened on Mg2(+)-Ca2+ exchange. Despite the clear distinctions in the thermodynamic features, the conformational changes of toad parvalbumin are essentially the same as those of the two isotypes of bullfrog parvalbumins on Ca2+ binding and Mg2(+)-Ca2+ exchange.