George E. Tarr

Reverse-Phase High-Performance Liquid Chromatography of Hydrophobic Proteins and Fragments Thereof'

Reverse-phase high-performance liquid chromatography (HPLC) resolution and recovery of cytochrome P-450 and bovine rhodopsin, both integral membrane proteins, and large peptides derived from P-450 LM2 were enhanced by utilizing ternary solvents. Surprisingly, most test materials eluted later in the gradient when using mixtures of acetonitrile and propanol in the mobile phase compared to using either solvent alone. Of the supports tested, the best recovery of hydrophobic cytochrome P-450 LM4 was experienced on the less retentive CN-bonded phase. Two alternate solvents for HPLC of polypeptides are proposed: (1) 0.02-0.1 M hexafluoroacetone/NH3, pH 7.2 for highly acidic peptides; and (2) 6 M formic acid/0.13 M trimethylamine, pH 1.5, vs 4 M formic acid/0.09 M trimethylamine in propanol for relatively insoluble peptides. Anomalous side reactions between formic acid and peptides can cause HPLC peak broadening, increased retention, and decreased resolution. These deleterious effects are thought to be due in part to formyl esterification of serine and threonine residues and appear to be reversible by aminoethanol treatment.

Manual Batchwise Sequencing Methods

Until recently the rate-limiting step in protein structural determination was the purification of peptides, so sequencing efforts focused on long degradations of large fragments. With the advent of HPLC methods the sequencer has been confronted with a host of suitable material mostly of small to moderate size. Improvements in procedure for automatic instruments like the spinning cup have extended their applicability, range and sensitivity, but have only slightly increased their speed: one day’s worth of HPLC runs can easily produce enough peptides to occupy an automatic sequenator for half a year.

Two-dimensional peptide mapping by reversed-phase column chromatography, applied to the sequence determination of cytochrome c from the wild type and a mutant of the butterfly, pieris brassicae

Abstract Two-dimensional peptide mapping has been very effective in the characterization of protein digests, particularly for the detection of small structural differences between homologous proteins. The classical thin-layer strategy, which exploits differences in charge and hydrophobicity, has been realized as a method based on reversed-phase high-performance liquid chromatography. An initial fractionation at pH 7.2 with 100 m M potassium phosphate, followed by chromatography with 0.1% trifluoroacetic acid, has been applied to chymotryptic digests of cytochromes c . The use of UV-transparent and (in the final stage) volatile solvents allows detection and rapid recovery of nanomole amounts of peptides suitable for sequence determination. As an example of the application of this method we report the comparison of two variants of cytochrome c from the butterfly, Pieris brassicae , one being the wild type and the other a spontaneous mutant isolated from a laboratory colony. The single residue difference was easily detected and identified.

Cloning and Sequence Analysis of the Structural Gene for the bc1-Type Rieske Iron-Sulfur Protein from Thermus thermophilus HB8

The structural gene encoding the Rieske iron-sulfur protein from Thermus thermophilus HB8 has been cloned and sequenced. The gene encodes a protein of 209 amino acids that begins with a hydrophilic N-terminus followed by a stretch of 21 hydrophobic amino acids that could serve as a transmembrane helix. The remainder of the protein has a hydrophobicity pattern typical of a water-soluble protein. A phylogenetic analysis of 26 Rieske proteins that are part of bc1 or b6f complexes shows that they fall into three major groups: eubacterial and mitochondrial, cyanobacterial and plastid, and five highly divergent outliers, including that of Thermus. Although the overall homology with other Rieske proteins is very low, the C-terminal half of the Thermus protein contains the signature sequence CTHLGC-(13X)-CPCH that most likely provides the ligands of the [2Fe-2S] cluster. It is proposed that this region of the protein represents a small domain that folds independently and that the encoding DNA sequence may have been transferred during evolution to several unrelated genes to provide the cluster attachment site to proteins of different origin. The role of individual residues in this domain of the Thermus protein is discussed vis-a-vis the three-dimensional structure of the bovine protein (Iwata et al., 1996 Structure4, 567–579).

C-Terminal Sequence of Proteins: Rapid Isolation and Edman Sequencing of C-Terminal Peptides from Digests

Methods for obtaining sequence information at the C-terminus of a protein, ideally as efficiently and conveniently as at the N-terminus using the Edman degradation, have been of perennial interest to protein chemists (see (2) for review). And now molecular geneticists would like information sufficient for construction of a probe, especially as the N-terminus is all too frequently blocked, and for verifying reading frames. The most popular current method for direct C-terminal sequencing is by carboxypeptidase, usually Y (3); besides the problems of phasing, widely varying residue- or sequence-specific rates, and unpredictable digestion conditions, the method apparently predicts a structure only rarely resembling the true sequence (in only 2 of 12 cases for cytochromes P-450 (4)). Of many chemical methods, only the 60-yr old thiocyanate degradation developed to practicality by Stark (5) has attained appreciable success, but even with repeated improvements (most recently, (6)) it seems not to have contributed any new sequence information. An alternative is the indirect “back door” approach of isolating from a digest of a protein that peptide(s) representing the C-terminal region, followed by Edman sequencing. This requires tagging the C-terminus before digestion and a convenient way of picking the tagged peptide from among the many others. For instance, all protein carboxyl groups, side-chain and C-terminal, could be reacted with an amine. Digestion restores one carboxyl group to each peptide except for the C-terminal one, thus permitting discrimination on the basis of this group (7), for instance by anion exchange chromatography; the unbound C-terminal peptide can be transferred directly to a reversed phase column for final purification. We have used an amidation procedure for several years that increases solubility and unfolds proteins to enzymatic digestion, and which gives peptides a definitive end-of-sequence signal (1,8), so our isolation of the C-terminal peptide extends this procedure.

Determination of the sequence which spans the beginning of the insertion region in Anacystis nidulans flavodoxin

Anacystis nidulans flavodoxin is one of the long chain flavodoxins (Mayhew & Ludwig, 1975). Comparisons of its structure with the structures of shorter chain species (main text: Ludwig et al., 1982) show that in A. nidulans flavodoxin most of the extra residues occupy a region adjoining the third helix and the fifth strand of parallel sheet. The sequences of peptides isolated after cyanogen bromide cleavage and after digestion with Staphylococcus aureus protease, reported here, fit into the electron density map of A. nidulans flavodoxin, starting near the middle of the third helix, and verify that the major insertion interrupts the fifth strand of parallel sheet.

Amidation of Protein Carboxyl Groups

One of the delights of working with proteins, as opposed to nucleic acids, is their great variation in properties. But for the protein sequencer concerned only with covalent structure, this is also a problem. For example, a protein which is acidic or easily precipitated by organic solvents may not elute from a reversed-phase HPLC column under the standard conditions of dilute acid and acetonitrile (MeCN) so effective in a majority of cases. Use of a neutral buffer (a UV-transparent, volatile one is available)1 alleviates problems due to acidity but aggrevates those due to precipitation. And many proteins that do behave well on HPLC may balk when treated with a protease: they are either insoluble in suitable buffers or present a refractory surface with susceptible cleavage sites buried. The usual stratagem is to include a denaturant, but the level of urea that trypsin and most other enzymes will tolerate is often insufficient to promote complete and repoducible digestion, and detergents — which also are not always effective — interfer with separations employing RP-HPLC. An alternative to adjusting conditions to try to accomodate each ill-behaved protein is to modify these proteins at the start in such a way that they are all alike in their properties and common problems are overcome. Ideally this modification should be specific, complete, stable, universally produced under gentle conditions, unrestrictive with regard to other manipulations, and advantageous during sequencing.

The Primary Structure and Posttranslational Modification of Human Hypoxanthine-Guanine Phosphoribosyltransferase

HPRT has been purified to apparent homogeneity from several different human tissues. Estimates of the apparent subunit molecular weight of the human enzyme have ranged from 26,000 to 24,500 (reviewed in reference 3). The human erythrocyte enzyme exhibits SUbstantial electrophoretic heterogeneity which is caused, in part, by at least two undefined posttranslational modifications (4,5). Insight into the mechanisms responsible for a deficiency of HPRT activity in man was recently provided by the isolation of structural variants of the human enzyme. Mutant forms of HPRT were purified from lymphoblasts (6) and erythrocytes (7) of 5 unrelated patients with a deficiency of enzyme activity. A detailed investigation of the function and subunit structure of these purified enzymes provided evidence for the existance of 5 unique structural variants.

Is cytochrome aa3 from thermus thermophilus a single subunit oxidase

Abstract A reliable procedure has been developed for the purification of the cytochrome c 1 aa 3 complex from the plasma membrane of T. thermophilus . The ratios heme C:heme A:Fe:C were found to be 1:2:3:2 confirming previous results, however, the molecular weight was found to be ∼92,000 rather than the ∼200,000 reported earlier [1]. Polyacrylamide gel electrophoresis under strongly denaturing conditions and high performance reverse phase liquid chromatography showed that cytochrome c 1 aa 3 is composed of only two subunits in 1:1 ratio. Both polypeptides have blocked N-termini. The smaller subunit (∼33,000) binds heme c and presumably no other metals. The larger subunit (∼55,000) is thus thought to contain the elements of cytochrome aa 3 and therefore be considered a single subunit cytochrome oxidase. The bacterial cytochrome c 1 aa 3 has been compared with beef heart cytochrome oxidase with a number of techniques including optical, EPR [1], Raman, MCD, and Mossbauer [2] spectroscopies. These experiments establish that the fundamental chemical properties of the redox centers are substantially similar in these two proteins. Cytochrome c 552 (from Thermus ), horse heart cytochrome c , and tetramethylphenylenediamine greatly stimulate the ascorbate oxidase activity of cytochrome c 1 aa 3 . This enhancement is characterized by a ‘high affinity’ component which results in only a small velocity increase and a ‘low affinity’ component which gives a large velocity increase. Very similar behavior has been previously observed with mammalian cytochrome oxidase [3]. Preliminary experiments show that vesicularized c 1 aa 3 is capable of proton pumping.