Towards a molecular understanding of phase separation in the lens: a comparison of the X-ray structures of two high Tc gamma-crystallins, gammaE and gammaF, with two low Tc gamma-crystallins, gammaB and gammaD.

gamma-Crystallins, although closely related in sequence, show intriguing differences in their temperature-dependent interactions: those that have a high or intermediate Tc for phase separation are cryoproteins whereas low Tc gamma-crystallins are not. To address the molecular basis of phase separation, X-ray crystallography has been used to define the structural differences between high and low Tc gamma-crystallins. A pre-requisite for this study was to clarify the assignment of bovine gene sequences to bovine gamma-crystallin proteins used for biophysical measurements. Based on nucleotide sequence analyses of gamma E and gamma F bovine crystallin genes, gamma F corresponds to the previously crystallised high Tc protein bovine gamma IVa and gamma E corresponds to the high Tc bovine protein fraction previously known as gamma IIIa. The gamma F sequence has enabled the completion of the refinement of the bovine gamma F crystal structure which shows that the molecule has an additional surface tryptophan explaining why gamma F has different spectroscopic properties from gamma B. A high Tc protein from rat lens, gamma E crystallin, has been crystallised and the X-ray structure solved at 2.3 A resolution. Comparison of the X-ray structures of two high Tc proteins, rat gamma E and bovine gamma F, with the structures of two low Tc proteins, bovine gamma B and bovine gamma D, shows that the main conformational change between high and low Tc proteins is in the cd surface loop of motif 3. All four structures have numerous ion pairs on their surfaces leading to a high surface charge density, yet with low overall charge. Comparison of the lattice contacts of the two high Tc proteins with the two low Tc gamma-crystallins indicates that these high Tc proteins utilise more amino-aromatic interactions such as between histidine and arginine. Comparison of the sequences of all the gamma-crystallins which have been characterised for phase separation temperature indicates that only residue Arg/Lys 163 uniquely distinguishes cryo from non-cryo gamma-crystallins and it is close to the altered surface loop. Although this region probably contributes to phase separation, Tc is likely to be a function of an overall global property that is responsive to overall charge distribution. Calculated dipole moments of native gamma-crystallins, low Tc gamma-crystallin sequences threaded into high Tc gamma-crystallin structures, and vice versa, show how both sequence and 3D structure contribute to this overall property. High Tc gamma-crystallins have on average higher Arg/Lys ratios and higher histidine content. It is hypothesised that this increases the proportion of surface static paired charged networks which thus reduces the repulsive hydration force and so increases the attractive interactions of the protein-rich phase in binary liquid phase separation.

Structure of bovine eye lens gammaD (gammaIIIb)-crystallin at 1.95 A.

The crystal structure of bovine lens gammaIIIb-crystallin at 2.5 A resolution previously reported was interpreted using a consensus sequence derived from related vertebrate sequences on the assumption that gammaIIIb-crystallin derived from the gammaC-crystallin gene. It has recently been shown that gammaIIIb is a product of the bovine gammaD gene. The structure of gammaIIIb has now been refined with the bovine gammaD sequence using new 1.95 A resolution synchrotron data. The crystallographic R factor was 20.4% for all 33 104 reflection data between 8.0 and 1.95 A measured at 277(1) K. The electron density fully supported the assignment of the gammaD sequence to gammaIIIb. The crystal belongs to space group P2(1)2(1)2(1) with two molecules of molecular mass 20 749 Da in the asymmetric unit in which 219 water molecules were located. The two-domain four-Greek-key motif highly symmetrical protein is very similar in structure to gammaB-crystallin (81% sequence identity). There is a single amino-acid deletion in gammaD in the linker region connecting the two domains. The intermolecular oganization in the crystal lattice is quite different from gammaB as a result of key mutations involving surface residues Leu51, Ile103 and His155. These point mutations will contribute to the intermolecular behaviour of the gamma-crystallins in the eye lens, where they are major components of the densely packed, high refractive index regions of the lens.

Expression of recombinant bovine gamma B-, gamma C- and gamma D-crystallins and correlation with native proteins.

Despite the use of bovine gamma-crystallins in numerous biophysical and chemical studies, characterization of these proteins at the molecular level is incomplete. Bovine lenses have at least six gamma-crystallin protein fractions currently assigned as gamma s/I, gamma A/IVb, gamma B/II, gamma C/IIIb, gamma D/IIIa and gamma E/IVa. A lack of primary sequence data for corresponding gamma-crystallin genes and proteins, however, has made these assignments tenuous. To clarify these assignments, we have over-expressed recombinant bovine gamma-crystallin proteins in Escherichia coli using complementary DNAs corresponding to gamma B-, gamma C-, and gamma D-crystallin genes. The recombinant crystallins were characterized by chromatographic and spectroscopic comparisons with native bovine crystallin fractions gamma II, gamma IIIa and gamma IIIb. The elution of recombinant gamma B and native gamma II proteins was identical on cation-exchange chromatography as expected; however, recombinant gamma C coeluted with gamma IIIa and recombinant gamma D co-eluted with gamma IIIb. Sequential Edman degradation through the first 29 residues of the native gamma IIIa and gamma IIIb polypeptides confirmed the colinearity of their sequences with those predicted from the gamma C- and gamma D-crystallin genes, respectively. Absorption and UV circular dichroism (CD) spectra of the recombinant proteins were almost identical to those of their native counterparts, indicating that the secondary and tertiary structures of the recombinant proteins were characteristic for gamma-crystallins. Based on these data the bovine gamma-crystallins proteins and genes are correlated as follows: II/gamma B, IIIa/gamma C and IIIb/gamma D. The assignment of gamma IIIb (previously characterized as having a low critical temperature for phase separation) with gamma D rather than gamma C proves an exception to the hypothesis that the gamma ABC-crystallin group is more resistant to phase separation than the gamma DEF group. These corrected assignments should provide a more substantial base for investigations of residues responsible for phase separation and other biophysical characteristics. Additionally, expression of recombinant gamma-crystallins with structures similar to native proteins may prove to be useful in probing specific structure-function relationships of the gamma-crystallins.

Nucleotide sequence of a bovine lens alpha A-crystallin cDNA.

We have determined the nucleotide sequence of a bovine lens alpha A2-crystallin cDNA clone, designated pBL alpha A2-1. The 793 bp cDNA insert contains coding information for the entire 173 amino acid alpha A2-crystallin polypeptide, as well as non-translated sequences located both upstream and downstream from the coding region. The coding sequences contained in pBL alpha A2-1 are at least 89% homologous with the corresponding sequences from other mammalian alpha A-crystallin genes, and are 78% homologous to the frog alpha A-crystallin coding region. In contrast, the downstream nontranslated sequences of the mammalian alpha A-crystallin transcripts show much greater sequence divergence, with the bovine sequences averaging 47% homology with the corresponding sequences from other mammalian species.

cDNA clones encoding bovine gamma-crystallins.

We have determined the nucleotide sequence of two bovine lens gamma-crystallin cDNA clones, pBL gamma II-1 and pBL gamma III-1. The 644 bp cDNA insert of pBL gamma II-1 contains coding information for the entire amino acid sequence of bovine gamma II-crystallin. The 497 bp cDNA insert of pBL gamma III-1 encodes a homologous but different gamma-crystallin polypeptide, and appears to lack the coding information for the C-terminal 17 amino acid residues. While the nucleotide and predicted amino acid sequences of the coding regions of the clones show a high degree of homology, the untranslated leader sequences are relatively dissimilar. The leader sequence of pBL gamma III-1 is strikingly homologous to a portion of a rabbit immunoglobulin alpha-heavy chain mRNA.

Nucleotide sequences that define promoters that are used by Bacillus subtilis sigma-29 RNA polymerase.

There are at least five different forms of RNA polymerase holoenzyme in Bacillus subtilis. These enzymes differ in their sigma subunit and their specificity for promoter utilization. One form of RNA polymerase (E sigma 29) that contains a 29,000 Mr sigma appears in B. subtilis about two hours after the initiation of endospore formation. The determination of the nucleotide sequences that govern utilization of promoters by E sigma 29 has been limited by the small number of cloned promoters that are recognized by E sigma 29. We have determined the nucleotide sequence of a recently isolated promoter (G4) that is used exclusively by E sigma 29 both in vitro and in vivo. The start-point of transcription was identified by S1 nuclease mapping and dinucleotide priming experiments and the probable promoter element was sequenced. We compared the sequence with that of six promoters that are used to varying degrees in vitro by E sigma 29 and found these sequences to be highly conserved at the -10 and near the -35 regions of these promoters. Single base substitutions were generated at positions -12, -15 and -36 of the G4 promoter and assayed for their influence on utilization by E sigma 29 in in-vitro competition experiments. The effects of these mutations in G4 on its use by E sigma 29 support a model in which E sigma 29 utilizes its cognate promoters by interacting with unique nucleotide sequences at the -10 region and near the -35 region of these promoters.

Promoter used by sigma-29 RNA polymerase from Bacillus subtilis.

Gene expression during endospore formation by Bacillus subtilis is controlled in part by a sporulation-induced form of RNA polymerase, E sigma 29. The determination of the nucleotide sequences that govern utilization of promoters by E sigma 29 has been limited by the small number of available promoters that are recognized by E sigma 29. In the present report we describe a promoter that is adjacent to the rrnB region of the B. subtilis chromosome and is utilized in vitro and in vivo by E sigma 29. S1 nuclease mapping and dinucleotide priming experiments have been used to determine the start point of transcription. The nucleotide sequences near the -10 and -35 region of this promoter, bvx, are conserved, and resemble sequences at these regions for other promoters that are utilized by E sigma 29.

Mutations that affect utilization of a promoter in stationary-phase Bacillus subtilis.

Transcription of the ctc gene in Bacillus subtilis is activated only after exponentially growing cells enter stationary phase. The promoter of the ctc gene is utilized in vitro by two minor forms of RNA polymerase, E sigma 37 and E sigma 32, but not by the most abundant form of RNA polymerase, E sigma 55. We have used the ctc promoter to direct transcription of the xylE gene on plasmid pLC4 and observed that xylE was expressed only in stationary-phase B. subtilis. We also have constructed a series of homologous plasmids that differ only by specific base substitutions in the ctc promoter. We observed that the base substitutions that affected utilization of the ctc promoter in vivo (xylE expression) were the same as those that we had previously shown to affect utilization of the promoter in vitro by E sigma 37 and E sigma 32. We conclude that it is likely that the ctc promoter is utilized in vivo by E sigma 37 or E sigma 32.

Promoter specificity of a sporulation-induced form of RNA polymerase from Bacillus subtilis.

The program of gene expression that underlies endospore formation by Bacillus subtilis may be controlled in part by a sporulation-induced form of RNA polymerase, E sigma 29. The nucleotide sequences of four promoters, which are known to be recognized by E sigma 29, are highly conserved at two regions, 10 bp and 35 bp upstream from the start point of transcription. We have used oligonucleotide-directed mutagenesis to construct several base substitutions in the ctc promoter from B. subtilis to test the role of the highly conserved sequences in utilization of the promoter by E sigma 29. In vitro transcription analysis demonstrated that the conserved nucleotides at positions -15, -14 and -12 affect the utilization of the promoter by E sigma 29. These and previous results support a model in which E sigma 29 recognizes its cognate promoters by interacting with nucleotides near the -10 and -35 regions. We also examined the effects of these base substitutions on utilization of the promoter by two other forms of RNA polymerase from B. subtilis, E sigma 37 and E sigma 32.

Covalent structure of human haptoglobin: a serine protease homolog.

The complete amino acid sequences and the disulfide arrangements of the two chains of human haptoglobin 1-1 were established. The alpha 1 and beta chains of haptoglobin contain 83 and 245 residues, respectively. Comparison of the primary structure of haptoglobin with that of the chymotrypsinogen family of serine proteases revealed a significant degree of chemical similarity. The probability was less than 10(-5) that the chemical similarity of the beta chain of haptoglobin to the proteases was due to chance. The amino acid sequence of the beta chain of haptoglobin is 29--33% identical to bovine trypsin, bovine chymotrypsin, porcine elastase, human thrombin, or human plasmin. Comparison of haptoglobin alpha 1 chain to activation peptide regions of the zymogens revealed an identity of 25% to the fifth "kringle" region of the activation peptide of plasminogen. The probability was less than 0.014 that this similarity was due to chance. These results strongly indicate haptoglobin to be a homolog of the chymotrypsinogen family of serine proteases. Alignment of the beta-chain sequence of haptoglobin to the serine proteases is remarkably consistent except for an insertion of 16 residues in the region corresponding to the methionyl loop of the serine proteases. The active-site residues typical of the serine proteases, histidine-57 and serine-195, are replaced in haptoglobin by lysine and alanine, respectively; however, aspartic acid-102 and the trypsin specificity, residue, aspartic acid-189, do occur in haptoglobin. Haptoglobin and the serine proteases represent a striking example of homologous proteins with different biological functions.

Canine haptoglobin: a unique haptoglobin subunit arrangement.

1. Isolated canine haptoglobin behaved identically to the alpha 2 beta 2 structure typical of human haptoglobin type 1-1 on alkaline polyacrylamide gel electrophoresis and on gel filtration. 2. In the presence of urea or sodium dodecyl sulphate canine haptoglobin dissociated into alpha beta subunits that separated into alpha and beta chains after reduction with 2-mercaptoethanol. 3. Compositional analysis identified one less half-cystine in canine alpha chain when compared to human alpha 1 chain. 4. These results provide evidence that there is no inter alpha chain disulphide in canine haptoglobin comparable to the alpha 1 20-alpha 1 20 disulphide in human haptoglobin that links the two alpha beta subunits.

Characterization of the cyanogen bromide fragments of the beta chain of human haptoglobin.

Characterization of the cyanogen bromide (CNBr) fragments of the beta chain of human haptoglobin revealed five major fragments resulting from cleavage of four methionyl residues. The fragments were isolated by gel filtration in guanidine-HCl on Sepharose 6B and Bio-Gel P10 and P60. Compositional analyses of the five cyanogen bromide fragments accounted for 248-253 amino acid residues in agreement with the number of residues determined for the intact beta chain. Most of the carbohydrate was attached to CNBr II. Automated amino-terminal sequence analysis and carboxyl-terminal hydrolysis with carboxypeptidase of the haptoglobin beta chain and cyanogen bromide fragments identified 139 residues, or about 55% of the beta-chain molecule. The placement of the fragments within the beta-chain molecule was established by sequence analysis of whole beta chain and a plasmin cleavage fragment. The position of CNBr V was confirmed by the absence of homoserine or homoserine lactone. Cyanogen bromide reaction of intact haptoglobin 1-1 resulted in the isolation of a beta-chain fragment, CNBr III, covalently attached to the intact alpha1 chain by a single disulfide bond. The beta chain was shown to have primary structural similarities to the chymotrypsin family of serin eproteases. Partial sequence analysis of CNBr V established the region which is comparable to the serine-195 active-site region: /Asp-Thr-Cys-Tyr-Gly-Asp-Ala-Gly-Ser-Ala-Phe/ (residues 189-199, chymotrypsinogen A numbering). The active-site serine-195 is replaced by alanine; however, the specificity residue of the trypsin-like enzymes, Asp-189, is preserved. Several minor cyanogen bromide cleavage products were also identified in yields of up to 15%. These minor cleavage products give evidence that tryptophanyl residues in proteins, or glycoproteins, are also susceptible to cyanogen bromide cleavage.