>I ' “r: 1 [flux-W? $103 smau. 'MEMBR MARY ! PR TIAL' F G553 G A ‘ OLE mu 0 3 \ ( 020m} R s fly £6.48 ... . m R: . - ” . . .2 Jugamnflw ”an? 2; A r. . .. Ia .r‘ 4..- .A...‘ . F . A tf:¥~..: ‘ . h . . t . . 11.9.llvur This is tocertify that the . thesis entitled PARTIAL PRIMARY STRUCTURE OF GAS VACUOLE NENBRANEB FROM MICROCYSTIS AERUGINOSA presented by Pamela J. Weathers has been accepted towards fulfillment of the requirements for Ph. D. Botany degree in ,L’Lir/QMZ J a F Major professor , DateJ/ 1:] 7r— 0-7639 is 800K BINDERY INC. ‘ll LIBQARV enemas l is [le111119051. mom“: ~ I \Afi l _____ ABSTRACT PARTIAL PRIMARY STRUCTURE OF GAS VACUOLE MEMBRANES FROM MICROCYSTIS AERUOGINOSA By Pamela J. Weathers The primary structure and hydrOphobic characteristics of the gas vacuole-membrane were investigated in order to determine if this simple membrane, containing only one protein, typifies inte- gral membrane proteins as defined by Singer and Nicholson (l7). Peptides resulting from trypSin digestion and N-Bromosuccini- mide treatment of gas vacuole protein were sequenced using a variety of methods. Sequence analysis from the amino terminus was performed using automated Edman degradation on a Beckman Sequencer Model 890C. Modification of the peptides to prevent their loss from the reaction cup was effected by making the carboxyl terminus more hydrophilic. This was accomplished by using either h-sulpho- phenyl isothiocyanate for lysine, or water soluble carbodiimide and naphthalene disulfonic acid for other amino acids. I improved the old carbodiimide procedure by using additional aliquots of carbo- diimide, especially for insoluble peptides. Sequence analysis from the carboxyl end of peptides was performed using carboxypeptidase C. Pamela J. Weathers All sequences are shown below and are preceded by their designated names: T2A3a: Ala-Val-Glu-Lys TlPlb: Tyr-Ala-Glu-Ala-Val-Gly-Leu-Thr-Glu-Ser-Ala-Ala-Val-Pro- (lS residues)-Arg-Tyr-Ala-Glu-Ala-Val-Gly-Leu-Thr-Glu-(Ser)- Ala-(Pro)-(Val)-Ala-Ala TlPla: Ser-Ala-Glu-Ala-Val-Gly-Leu-Thr-Glu-Val-(|le)-(Ala)- (x number of residues) T2A2: Gly-lle-Val-lle-(Asp)-(Ala)-Ala-Arg T2A3b: lle-Leu-Asp-Lys T2Ah: Lys NlA: Ala-Glu-Ala-Val-Gly-Leu-Thr-Glu-(Ser)-Ala-(Pro)-(Val)-Ala- Ala NPT: Ala-Val-(Val)-(Val)-Leu-Val-(Val)-|le-(lle)-Leu-(Leu)-Ala- (Leu)-(Val)-(lle)-(x number of residues) Peptide T2A3a is the amino terminus of the protein; peptide TlPlb, the carboxyl terminus. Peptide NPT and all others reside somewhere between the two terminal peptides. Two aspects of the sequence analysis are intriguing. Firstly, the sequence of peptide NPT is one of the most hydrophobic sequences of a protein yet described; such a long, aliphatic stretch is rare. Based on the amino acid sequence of a protein, current methods (32) permit determination of the regions of secondary structure (helix, sheet, turn) in a protein with 85% certainty. Application of these methods shows that the amino acid sequence of the NPT peptide would equally favor either helix or sheet formation. Pamela J. Weathers The second important feature of the primary structure of the gas vacuole protein is the presence of the thrice repeating octa- peptide: Ala-Glu-Ala-Val-Gly-Leu-Thr-Glu A determination of the secondary structure, as above, indicates that this sequence strongly favors helix formation. It is generally inferred (85) that the presence of a repeating sequence in a poly- peptide suggests its use as a structural building block. Based on the presence of peptide NPT, a functional model for the gas vacuole membrane was proposed. Gases pass through the mem- brane by diffusion. Since many diatomic gases are apolar, passage through an apolar milieu would provide a path of least resistance. Such an apolar milieu would exist in the aliphatic, amino terminal portion of the peptide NPT. The assumption is made that a subunit substructure exists in the membrane. To allow gas to move in and out without a conformational change in the membrane (for which there is no evidence at present), either the gas must pass through the intermolecular space in a protein or through pores. The lining of either of these passageways would be the aliphatic portion of the NPT peptide. Large and/or polar molecules would be restricted from passing through this area based on their size (at least 3 A), and charge. The presence of the peptide NPT internal to the amino and carboxyl termini of the gas vacuole protein, provides the protein Pamela J. Weathers with an amphipathic nature. Based on this and a comparison of the relative polarities of gas vacuole protein to other integral membrane proteins, it is concluded that the gas vacuole protein is a true membrane protein of the integral type. Perhaps, this membrane is even a prototype for more complex membranes. PARTIAL PRIMARY STRUCTURE OF GAS VACUOLE MEMBRANES FROM MICROCYSTIS AERUGINOSA By Pamela J. Weathers A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Botany and Plant Pathology l97h DEDICATION to my husband, Larry, and our Oakhill families ACKNOWLEDGMENTS The author wishes to express her gratitude to her major professor, Dr. Michael Jost, for his guidance throughout these investigations. Special appreciation is extended to Dr. Derek T.A. Lamport for his extensive advice, use of his laboratory equipment, and encouragement throughout these studies. The constructive criticism and advice of the committee members Dr. Alfred Haug and Dr. Robert Bandurski is also appreciated. The author also wishes to thank Dr. Judith Foster of the Boston University School of Medicine for her helpful discussions and guidance in sequencing procedures. The technical assistance of Nell Brittain, Ray Sculley, Laura Katona, and Mary Shimamoto is gratefully acknowledged. This work was supported under Contract No. AT(ll-l)-1338 with the U.S. Atomic Energy Commission. TABLE OF CONTENTS LIST OF TABLES. . . . . . . . . . . . . . . . . . . . L'ST OF FIGURES O O O O O O O O O O O O O O O O O O 0 LIST OF ABBREVIATIONS . . . . . . . . . . . . . . . . INTRODUCTION l.l Membrane Structure: Fluid Mosaic Model . . I.2 The Nature of Membrane Proteins . . . . . . l.3 Integral Membrane Proteins. . . . . . . . . I.h Structural Analysis of Integral Membrane Proteins. . . . . . . . . . . . . . . . . Description of the Membrane System Used in this Study: The Gas Vacuole Membrane . . Implications of Sequence Data . . . . . . . Techniques Available for Sequence Analysis. Objectives of This Study. . . . . . . . . . did d O O mNCh U1 MATERIALS AND METHODS Culture, Harvest, and Lysis . . . . . . . . Purification of Gas Vacuole Membranes . . . Maleylated Tryptic Peptides . . . . . . . . N-Bromosuccinimide Peptides . . . . . . . . Digestion of the NPT Peptide by Thermolysin Partial Acid Hydrolysis of the GVP and Peptides. . . . . . . . . . . . . . . . . Recovery of Peptides (Balance Sheet). . . . Purity of Peptides. . . . . . . . . . . . Sequence Analysis of Peptides by Carboxypeptidase C. . . . . . . . . . . . 2.l2 Amino Acid Analysis . . . . . . . . . . . . 2.l3 Automated Sequencing Methods. . . . . . . . 2.lh Identification of Amino Acid Derivatives from Sequence Analysis. . . . . . . . . . o —"'"‘\O CDNO‘UHJr'WN— -'O NNNNNNNN O NNN 0. Estimate of the Purity of Fraction E . . . . . Preparation and Separation of Tryptic Peptides. Page vi vii \ONNU’T U1 WN‘ ID ID l2 I2 l5 2] 22 22 22 25 25 25 36 RESULTS 0 O Nd wwwwwwwww WW Purity of Gas Vacuole Membranes. . . . . . Amino Acid Composition of Gas Vacuole Protein. . . . . . . . . . . . . . . . Sequence Analysis of GVP . . . . . . . . . Estimation of the Number of Tryptic Peptides Separation of the Tryptic Peptides . . . . Sequence Analysis of the Tryptic Peptides. Maleylated Tryptic Peptides. . . . . . . . Dilute Acid Hydrolysis of GVP. . . . . . N-Bromosuccinimide (NBS) Peptides. . . . . 0 Sequence Analysis of the NBS Peptides. . . I Improved Peptide Modification Methods for Sequencing . . . . . . . . . . . . . DISCUSSION h.I h.2 FF? rs \lO‘UW kw General Problems Encountered in Purification, Separation, and Sequencing of GVP Peptides . An Improved Method for Reduction of Extractive Losses of Peptides . . . . . . . . . . . . . Molecular Weight Determination of GVP. . . Sequence Analysis of Intact GVP and Alignment of Peptides. . . . . . . . . . Discussion of the Peptide NPT. . . . . . . A Repeating Octapeptide in the GVP . . . . Implications of the Amino Acid Composition of GVP and Its Relationship to Other Integral Membrane Proteins . . . . . . . Tentative Molecular Model for the Function of the Gas Vacuole Membranes . . . . . . BIBL'OGRAPHY O O O O O O O O O O O O O O O O O O O O Page 37 37 37 40 no 56 67 68 68 77 92 95 96 97 98 100 IOS 108 III IIA Table LIST OF TABLES Methods and Results of Sequence Analyses of Gas Vacuole Protein (Fraction E) . . . . . . . . . . Improved Peptide Modification Procedure. . . Purity of Cell Free Fractions Containing Gas Vacuole Membranes. . . . . . . . . . . . . . Amino Acid Composition of GVP. . . . . . . . Amino Acid Composition of the Purified Tryptic Peptides o o o o o o o o o o o o o 0 Recovery of Tryptic Peptides . . . . . . . . Sequence Analyses of the Tryptic Peptides from Amino Acid Composition of the NBS Peptides . Recovery of NBS Peptides . . . . . . . . . . Sequence Analyses of the NBS Peptides from GVP Recovery of the Tryptic Fragments of NPT . . Recovery of Fragments from Peptide NPT . . . Amino Acid Analyses of the Peptides Observed after Dilute Acid Hydrolysis of NPT. . . . . Relative Content of Polar, Intermediate, and Apolar Amino Acids for Some Membrane-Bound Proteins . . . vi Page 27 35 38 39 SI 55 60 73 76 80 8h 90 9i IOA LIST OF FIGURES Figure IO. II. 12. I3. Fractionation Scheme of the Cell Free Preparation of the Gas Vacuole Membranes . . . . . . . . . . . Fractionation Scheme of the Preparation and Purification of the Tryptic Peptides from Gas vaCUOEe Protein (GVP) O C O O O O O O I C O C O O O Fractionation Scheme of the Preparation and Purification of the Lysine Blocked Tryptic Peptides of GVP. . . . . . . . . . . . . . . . . . Maleylation. . . . . . . . . . . . . . . . . . . . Fractionation Scheme of the Preparation and Purification of the N-Bromosuccinimide (NBS) Peptides Of GVP. O O O O O I O O I I O C O O O O O Cyanoethylation. . . . . . . . . . . . . . . . . . The Edman Degradation for Sequence Analysis Of PO‘ypeptideSo O O O O C O I O I O O O I O O I 0 Operational Scheme for Peptide Program #02IS72 for Beckman Sequencer Model 890C . . . . . . . . . Peptide Modification Using Sulphophenyl Isothiocyanate (SPITC) . . . . . . . . . . . . . . Peptide Modification by Carbodiimide and Amino Naphthalene Sulfonic Acid. . . . . . . . . . . . . Minimum Estimation of the Number of Tryptic Peptides O O I I O O O I O O O O I O O O O O O O 0 Separation of the Tryptic Peptides on Sephadex 6-25 in 25% Formic Acid (2.4).. . . . . . . . . . . . . Separation of Fractions TI and T2AVI on SP Sephadex C-25-l20 . . . . . . . . . . . . . . . vii Page II 13 I6 18 20 2h 29 30 32 3h 42 Ah #6 LIST OF FlGURES--continued Figure l4. Separation of Tryptic Fraction T2 on Aminex A-S. I5. Elution Profile of Tryptic Fragment TIPI on Sephadex 6-25 in 25% Formic Acid . . . . . . l6. EIution Profile of Tryptic Fragment TIPI on Sephadex G-25 in 0.1 M Acetic Acid . . . . . . . . I7. Release of Amino Acids from Peptide TIP2 by Carboxypeptidase C . . . . . . . . . . . . . . . . I8. Release of Amino Acids from Peptide TlPlb by Carboxypeptidase C . . . . . . . . . . . . . . . . I9. Release of Amino Acids from Peptide T2A2 by Carboxypeptidase C . . . . . . . . . . . . . . . . 20. Separation of the NBS Peptides on Sephadex 6-25 in 0.] M Acetic ACid. O C O O O O O O O O O O O O O 0 2I. Composite EIution Profiles of the NBS Fractions NI, N2, and N3 on Sephadex G-25 in 25% Formic Acid 22. Separation of Fractions (Nla + Nlb) on Sephadex 6-25 in 09‘ M ACGLIC ACido o o o o o o o o o o o o 23. Release of Amino Acids from Peptide NIA by Carboxypeptidase C . . . . . . . . . . . . . . . . 2h. Elution Profile of the Tryptic Peptides TNPTa-e Of NPT O O O O O O O O O U C O O O O C O O O O I O 25. Separation of NPT Peptides Released by Thermolysin . 26. Separation of the Peptides DNa-d Released by Dilute Acid Hydrolysis of Peptide NPT. . . . . . . 27. Peptide Modification for Improved Sequencing . . . 28. Alignment of the Tryptic and NBS Peptides in the GVP. O O O I O O O O O O O O O O O O O O O O O O 0 viii Page A8 50 5h 59 62 65 70 72 75 79 83 87 89 9h I02 GVP CD EDC ANS SPITC PTH TMS NBS DNS DMAA Tris SDS DAP I LIST OF ABBREVIATIONS gas vacuole protein circular dichroism infra red analysis N-ethyl, Nl-(-3-dimethylamino prOpyI carbodiimide) HCl 2-amino-l, 5 naphthalene disulfonic acid h-sulphophenyl isothiocyanate phenylthiohydantion trimethylsilylated derivative N-Bromosuccinimide dansylated derivatives dimethylalIyIamine-trifluoroacetic acid in pyridine-water tris (hydroxymethyl) aminomethane sodium dodecyl sulfate dipeptidyl aminopeptidase I INTRODUCTION l.l Membrane Structure: Fluid Mosaic Model. Biological membranes play a crucial role in many cellular phenomena. Concern- ing the spatial arrangements of membrane constituents, relatively little is known. Therefore, models have been proposed. One of the most favored is the fluid mosaic model proposed by Lenard and Singer in I966 (I). This model accepts the basic lipid bilayer as origin- ally prOposed by Danielli, Davson and Robertson (2,3) with proteins inserted into the bilayer in a mosaic fashion. Some proteins may penetrate partially, whereas others may span the entire bilayer. In addition, the model suggests that proteins can migrate laterally in the membrane depending on the fluidity of the bilayer. Evidence for each of these basic tenets of the model comes from various membrane systems. Although there has always been experimen- tal data supporting a lipid bilayer (I, A, 5, 6), its existence only recently has been demonstrated unequivocally by X-ray diffraction (7) for Acholeplasma membranes. The concepts of fluidity and lateral mobility of lipids and proteins In the membrane have been discussed recently in a review by Singer (8). It has, for example, been shown for membranes of ‘A. laidlawii that temperatures at or near the phase transition temp- erature of the membrane lipids can affect the surface pattern of associated membrane proteins. This and other evidence (9, IO, II) also establishes that membranes contain proteins which are partially or wholly intercalated into the lipid bilayer. Although a fluid mosaic model is generally accepted in an architectural sense, the nature of the interaction occurring between the lipid and protein moieties of the membrane is still under consid- eration. However, covalent linkages are generally excluded as a major type of interaction between lipid and protein within the mem- brane (l2). On the contrary, these interactions must be weak and can be mediated by charge-charge interactions, hydrogen bonding, van der Waals interactions, and hydrophobic bonding (l, IZ, l3, lh). 1,2, The Nature of Membrane Proteins. Analysis of the secondary structure of the proteins of intact membranes shows that they contain about AO% helix which suggest that the proteins associated with the membrane are largely globular in shape (I2, l5, l6, l7). However, since these data were obtained using circular dichroism (CD), Optical anomalies arising from light scattering by particulate systems pre- clude exact calculation of the amount of secondary structure in the membranes (l7). There is a large body of evidence which suggests that proteins of biological membranes are asymmetrically distributed across a membrane. No evidence for a corresponding distribution of lipids has yet been found (8). Consequently, only examples of protein asym- metry will be mentioned here. These examples are membranes of Sarcoplasmic reticulum (II), the cytochrome b 5 reductase proteins of microsomal membranes (l8, l9), and glyc0phorin and MAD-cytochrome b5 of the erythrocyte membrane (20). Singer and Nicolson (l7) have categorized membrane proteins as either peripheral or integral. Peripheral proteins require only mild treatment to dissociate them from membranes; they dissociate free of lipids; and they are rela- tively soluble in neutral aqueous buffers. Integral proteins, which make up the major portion (more than 70%) of the proteins in most membranes, require very drastic treatments to dissociate them from 'membranes; once isolated, they may remain associated with lipids; and if free of lipids, they are usually highly insoluble in neutral aqueous buffers. Only the integral membrane proteins are critical to the structural integrity of the membrane (l7). In addition, there is no convincing evidence that only one predominant type of membrane structural protein exists (l7, 2l). Because peripheral membrane proteins are less involved in membrane structure than integral membrane proteins, they will not be discussed further. I.3 Integral Membrane Proteins. One of the most striking features of all integral membrane proteins studied, so far, is that they are extremely hydrOphobic. The majority of integral mem- brane proteins have a polarity,as determined by their amino acid composition, which is significantly lower (on the average lO-I9%) than that of most soluble proteins (22). However, hydrophobicity is not always due to a higher percentage of apolar amino acids. For example, penicillinase, a lipoprotein, is considered an integral membrane protein, yet its hydrophobic properties are due to the lipid moiety rather than an abundance of hydrophobic amino acids (23). Little is known about the role of hydrophobic bonds (2h), interactions involving nonpolar amino acids with nonpolar side chains. Several theories, though, have been espoused relating hydrOphobicity to thermal stability. Scheraga g£_gl. (25) and Bigelow (26) suggested that hydrOphobic bonds might play a crucial role in the thermal stabil- ity of proteins. However, studies of proteins from thermophilic organisms (27) and calculation of the hydrophobicity index, based on amino acid composition, relative to the thermal stability of a variety of proteins (28, 29) have demonstrated that no apparent relationship exists. Thus, the extreme hydrophobicity of integral membrane proteins is probably not due to the need for thermal stability of the membrane. Rather, hydrOphobicity is probably the mechanism whereby the protein can successfully merge with or span the lipid bilayer (8, l5, l7). To incorporate with an apolar matrix, integral membrane proteins must be amphipathic, i.e., they must contain polar and apolar regions in their primary structure. For example, glycophorin from the erythro- cyte membrane has been recently shown to have an amphipathic nature whereby the hydrophobic mid-portion of the protein spans the lipid matrix such that the amino terminal portion of the molecule is ex- posed to the exterior of the cell, while the carboxyl terminal seg- ment extends towards the interior of the cell (20, 30, BI). The cyto- chrome b5 protein has a very hydrophobic peptide, which is firmly bound to the lipid bilayer of the membrane (l8). However, no evidence exists to suggest that this protein spans the lipid bilayer. Rather, it apparently is anchored in the membrane by its hydrOphobic peptide (19). This type of partial penetration of the hydrophobic portion of a protein is also found for NAD cytochrome b5 reductase (l9). These observations suggest that hydrophobic bonding must play a crucial role in lipid-protein interactions (8, l2, l3, IA, l5, I7, 20). l.h Structural Analysis of Integral Membrane Proteins. Analysis of the primary structure of many integral membrane proteins should determine if these proteins are commonly amphipathic. The approximate secondary structure of the apolar regions can be deter- mined using eith CD or recent predictive methods requiring only an amino acid sequence (32). The latter method, devoid of computer calculations, makes use of empirically derived rules for predicting the initiation and termination of helix and sheet regions in proteins with 85% certainty. Other methods for predicting the occurrence of secondary structure in proteins had been attempted in the past, however, these were primarily concerned with helix determination and their reliability was poor (33, 3h, 35, 36, 37, 38). Nevertheless, the ability to predict the secondary structure of proteins will provide a basis for devising molecular models for the hydrophobic interactions between integral membrane proteins and the lipid bilayer in a membrane. l.5 Description of the Membrane System Used in This Study: The Gas Vacuole Membrane. Since the determination of secondary structure in proteins relies on the primary structure of the molecule, sequence analysis of integral membrane proteins is necessary. In this thesis, attention has been focused on one particular membrane system, the gas vacuole membrane from Microcystis aergginosa Kuetz. emend Elenkin. This membrane was chosen because it is simple. It contains only one protein species, no carbohydrate, and no lipid (39). In addition, the protein contains a high proportion of hydrOphobic, but no sulfur- containing amino acids (39). Yet, the gas vacuole membrane functions as a semi-permeable barrier whereby only gases are permitted to dif- fuse through the membrane (AO). Two functions have been attributed to the membranes: providing bouyancy to the cells (h0,4l), and light shielding of the photosynthetic lamellae (AZ). The molecular weight is reported as lh,OOO (39), however, there remains some confusion about this value. Preliminary X-ray data have suggested that a repeating “unit cell” (molecular weight 7,800) might exist (Al). On the other hand, polyacrylamide gel electrophoresis at pH 2 and A indicated one lh,300 molecular weight species (39), whereas SDS disc gel electrophoresis at pH 8.5 indicated a 2l,500 molecular weight Species (#3). Together, these results could indicate that a 7,000 molecular weight species exists which aggregates during polyacrylamide gel electrophoresis. Unpublished X-ray data have indicated an asymmetric electron density profile (hl). This was interpreted to mean that hydrophobic amino acids, which are less electron-dense than the hydrophilic amino acids, predominate on the inner (gas-facing) surface, whereas the hydrophilic amino acids predominate on the outer (cytoplasm-facing) surface (Al). In addition, several prominent wide-angle reflections were observed which characterize an extensive B-structure: a feature supported by CD analysis (Ah). Based on this, it was proposed that two layers of cross B-structure along each rib might account for the rigidity of the membrane. Blaurock and Stockenius (Al) have also suggested that this cross B-conformation would be such that along any rib, molecules would be joined by hydrogen bonding between the backbones of the chains. This last conjecture on the presence of intrarib hydrogen bonding is independently supported by others (A3, A5). No X-ray data are presented suggesting either the pres- ence or absence of helix in the membranes. However, IR (A6) and CD (AA) analyses have indicated that some helix is present. Analyses of the GVP from other species have been reported (A7, A8). Chemical information beyond a molecular weight, and amino acid composition remains incomplete. For a more thorough discussion on the structure and function of gas vacuoles, see Walsby's review (Al). l.6 Implications of Sequence Data. Sequence analysis, though often difficult, can provide not only the primary but also secondary structure of a protein. In addition, accumulation of the primary structures of many related proteins e.g. the cytochromes c, provides information for devising phylogenetic relationships among different organisms (A9). The genetic and evolutionary significance of changes which occur in the primary and secondary structures of various pro- teins is discussed in several reports and will not be considered here (A9, 50, SI, 52). l.] Techniques Available for Sequence Analysis. Hydrophobic proteins are especially difficult to sequence due to their insolubility; up to now, only one hydrophobic protein, glyc0phorin, has been sequenced (30). If peptides can be generated from a protein for sequence 8 analysis, many methods are available to assist in the determination of the actual structure. These methods can be enzymatic, chemical or physico-chemical. Enzymatic methods are those which use either carboxpeptidases A, B, or C (of which type C is the best (53)L or aminopeptidases. These enzymes sequentially degrade a peptide from either the carboxyl or amino terminal end. Providing the sequence contains no repeating residues and the peptide is very pure, these methods work. Another enzymatic approach to sequence analysis makes use of dipeptidyl aminopeptidase I (DAPI) which essentially cleaves peptides from their amino terminus into dipeptides (5A). Removal of the amino terminus of the peptide and subsequent retreatment with DAP I produces an overlapping set of dipeptides. The dipeptides are then separated and sequenced using chemical methods or mass spectroscopy (55). Major disadvantages of the method are the difficulty of separation of dipeptides, and limitation on the size of peptide to be sequenced (maximally 20-30 residues) (5A). Chemical methods of sequence analysis are extensive and only the most recent techniques will be mentioned here (56). All of the most effective methods use the Edman degradation (57). Problems arise, however, in losses of the peptides being sequenced to the or- ganic wash phase, termed washout. Techniques have been devised to overcome this problem by making the carboxyl terminus of the peptide more hydrOphilic. If the peptide in question is a tryptic peptide, those terminating in lysine can be modified by using A-sulfophenyliso- thiocyanate which is specific for primary amino groups (58,59). For peptides which terminate with other amino acids, modification is effected with 2-amino- l, 5-naphthalene disulfonic acid (ANS) using a water soluble carbodiimide (EDC) as the coupling agent (60). The primary amino groups are kept protonated so the ANS, being essentially uncharged, serves as the nucleophile. Polymerization of the peptides is, thus, avoided while amide linkages are formed at the free carboxyl groups. Utilization of these methods, which I have modified, reduces losses of peptides to washout. Another method which is relatively new is solid phase sequencing (61, 62). This procedure involves attaching the carboxyl terminus of the peptide to an insoluble matrix which then permits sequence analy- sis without the problem of washout. However, only recently has this technique been developed for practical use. l.8 Objectives of this Study. The basic objectives which I wished to accomplish in this study were threefold.» First, I would like to establish whether or not the gas vacuole protein is, indeed, a true membrane protein. A second objective, which might be expec- ted to follow from amino acid sequence data, is the development of a molecular model for the diffusion of gases into the membrane. My last objective concerns the hydrophobicity of the membrane. An analysis of the primary structure of glycophorin (30) has provided valuable information on hydrophobic proteins while generating the potential for studying hydrophobic interactions between integral mem- brane proteins and lipids. It is the purpose of this investigation to increase our knowledge of the hydrophobic properties of the GVP so that it might be used by other workers in the field as a model system in which hydrophobic interactions can be studied. l0 MATERIALS AND METHODS 2:] Culture,gHarvest,gandiLysis. The strain NRC-l of the Cyanophyte, Microcystis aeruginosa Kuetz. emend. Elenkin, obtained from P. R. Gorham,was grown in ASM-l medium (63) as described by Jones and Jost (39) to a final density of IO7 cells/ml. After incu- bation with benzylpenicillin, the cells were collected with an Amicon CH3 ultra-filtration concentrator fitted with an HICPlO fiber filter. The concentrate was then osmotically lysed (39). .2;2, Purification of Gas Vacuole Membranes. The procedure used is summarized in Figure l. Immediately after lysis, the lysate, fraction A, was centrifuged in a Sorvall Type SS-3A angle head rotor at 270 x g overnight; for the last I5 minutes of the run the rotor was accelerated to 3,000 x g. The pale blue-green layer, fraction 8, at the meniscus of the centrifuge tubes was drawn off and subse- quently filtered through a series of Millipore filters of decreasing pore size: 5 pm, 1.2 pm, 0.8 pm, 0.65 pm, O.A5 pm. The filtrate, fraction C, was then processed on a 5 x 50 cm column of Sepharose AB in 0.0l M Tris-HCl buffer, pH 7.5 containing 0.0l M NaN The milky- 3. white fraction 0 from the column was layered in 2.5 ml portions onto 9.5 ml of O.A M NaCl in centrifuge tubes and spun for 30 minutes at 200,000 x g in a Spinco SW-Al rotor. The narrow white band at the meniscus of the tubes was removed and stored at AOC in the presence of 0.0l M NaN3. The concentration of this purified gas vacuole membrane protein,fraction E,was determined by the absorbancy at A00 nm (A5). ll Intact Cells Penicillin Treatment Glycerol Induced Osmotic Shock Lysate (Fraction A) Flotation (270 x g) Pellet + Supernatant Pale Blue-green Band at Meniscus (Fraction B) ”'5 a'd Millipore Filtration I 771 Residue Filtrate (Fraction C) Discard Sepharose AB White Eluant (Fraction 0) Flotation (200,000 x g) White Band at Meniscus Remains (Fraction E) Discard Figure l. Fractionation Scheme of the Cell Free Preparation of Gas Vacuole Membranes. l2 2:} ,gstjmate of the Purity of Fraction E. Cells from a late log phase culture were innoculated into 500 ml of ASM-I medium to a final density of 6 x I06 cells/ml. The MgSOu in the medium was replaced by MgClZ. To this was added 0.0A mM NaZBSSOI+ (specific activity 27.8 mC/m mole) and O.l6 m M Naz3 SO“. Cells were then grown, harvested, and purified as described previously. At each step of the purification procedure 0.5 ml aliquots were assayed for protein by the method of Lowry g£_§1, (6A), and for radio- activity. The scintillation fluid used consisted of 907 ml dioxane, loo 9 naphthalene, and 5 g PPO. 2.A Preparation and Separation of Tryptic Peptides. The procedures used are outlined in Figure 2. Dialysis tubing was pre- washed with 88% formic acid. One volume of fraction E containing l00 mg of gas vacuole protein (GVP) was dialyzed against two volumes of 88% formic acid at room temperature. This acid denatured protein was then dialyzed exhaustively against deionized distilled water un- til the washing fluid reached pH 3. The acid precipitated protein in 0.002 M CaClz at pH 7.8 was then trypsinized for A8 hours at 250 C with two aliquots each of A mg of trypsin (specific activity I95 U/mg, Worthington). The first aliquot of trypsin was added at the begin- ning, the second, 18 hours after the start of the digestion. The pH of the reaction mixture was maintained with O.l N NaOH delivered by a pH stat or manually. At the end of trypsinization, the volume of the incubation mixture was reduced to dryness by flash evaporation at A00 C. An aliquot of the digestion mixture was dansylated (65), l3 Intact GVP (Fraction E) Dialysis Against 88% Formic Acid (2X) Acid Denatured GVP Dialysis Against Distilled H 0 2 Acid Precipitated GVP Trypsin Digestion Tryptic Peptides ‘Sephadex G-25 25% Formic Acid i__ Fraction Tl Fraction T2 Aminex A-5 Pool with Pyridine-Acetate T2AV I I I I I T2AV T2Al T2A2 T2A3 TZAA Tl + T2AV SP Sephadex C-25-l20 Formic Acid pH 2.3 0.05 M - 0.7 M NaCl Lys T2A33 Ala-Val-Glu-Lys T2A3b lle-Leu-Asp-Lys F’ 1 TIPI TlP2 Gly-lle-Val-Ile-(Asp)-(Ala)-Ala-Arg Sephadex G-25 Sephadex G-25 25% Formic 25% Formic Acid Acid TlP2 Desalted I_______1 TlPla TlPlb Tyr-Ala-Glu-Ala-Val-Gly-Leu-Thr-Glx-(Ser)-Ala-(Pro)-(Val)-Ala Ser-Ala-Glu- Ala-Val-Gly- Sephadex G-25 Leu-Thr-Glu- Val-(Ile)-(Ala)- 0.l M Acetic Acid (x residues) TlPlb Desalted Lys Tyr-Ala-Glu-Ala-Val-Gly-Leu-Thr-GIx-Ser-Ala-AIa-Val-Pro (l5 residues)-Arg-Tyr-Ala-Glu-Ala-Val-Gly-Leu-Thr-Glx- (Ser)-Ala-(Pro)-Ala-Ala- Figure 2. Fractionation Scheme of the Preparation and Purification of the Tryptic Peptides from Gas Vacuole Protein (GVP) IA acid hydrolyzed, and electrophoresed at pH l.9 (formic-acetic acid) at 5 kV for I35 minutes. The undansylated material was suspended in 2 ml of 25% formic acid and fractionated on a column (2 x IOO cm) of Sephadex G-25 (fine) in 25% formic acid (20). The different fractions were monitored for the presence of protein either at 280 nm or with ninhydrin at 570 nm. The latter assay was done with IS pl aliquots in a Technicon Autoanalyzer. The major fractions from the Sephadex column were designated TI and T2. Fraction Tl plus the fraction T2AV (See Fig. 2) were pooled and purified on a column (2 x 25 cm) of Sulpho Propyl Sephadex C-25-l20 (SP Sephadex fine) using a biphasic 0.05-0.7 M NaCl gradient in a formic acid-water mixture at pH 2.3. T2 was subsequently separated on a column (I x 25 cm) of Aminex A-5 (Bio-Rad) using a pyridine-acetate buffer system (66). The chambers of the gradient maker were loaded as follows: pyridine acetic acid tong vol. in chamber (ml) Chamber I had 0.068 M 2.7 70 2 0.068 M 2.7 70 3 O.l77 M 3.l 70 A O.l77 M 3.l 7O 5 O.l77 M 3.l 7O 6 O.l77 M 3.l 7O 7 2.0 M 5.0 70 8 2.0 M 5.0 70 IS The eluant profiles from both the Aminex A-5 and the SP Sephadex columns were monitored for the presence of protein as described above. Fractions from the SP Sephadex column were desalted over Sephadex G-25 (2 x I00 cm) in either 25% formic acid or 0.I M acetic acid at pH 2.87. 2.5 Maleylated Tryptic Peptides. The procedure used is outlined in Figure 3 under A. Fraction E was acid denatured as described above. The.c-amino groups of the lysines of the water- suspended GVP were maleylated according to Butler g£_gl. (67) (See Figure A) and then trypsinized as described above. Subsequent chromatography was over a Sephadex G-25 column (2.5 x I00 cm) in 0.l M acetic acid at pH 2.87 which effectively demaleylated the peptides. The different fractions were monitored for the presence of protein with ninhydrin at 570 nm and the TI and T2 fractions were pooled and further separated on Aminex A-5 and/or SP Sephadex using the previously described gradients for separation of the tryptic peptides. An alternative method of preparing maleylated tryptic peptides was that of irreversibly denaturing the protein as shown in Figure 3 under B. This was accomplished by heating the acid denatured pro- tein to l05o C for 30 minutes and maleylating the material as des- cribed above. Trypsin treatment was performed as described above and the digest was chromatographed over Sephadex G-25, G-SO, and G-75 in 0.0] M Tris-HCl buffer at pH 8.5. Fractions brought to pH 5 were analyzed with ninhydrin at 570 nm. I6 Intact GVP (Fraction E) Dialysis Against 88% Formic Acid (2X) Acid Denatured GVP Dialysis Against Distilled H20 Acid Precipitated GVP I Maleic'Anhydride (A) Maleylated GVP Trypsin Digestion Maleylated Tryptic Peptides Sephadex G-25 I l Tl T2 (Continued as in Figure 2) l Boiled Vacuoles (8) Acid Precipitated GVP (cooled) Maleic Anhydride Maleylated GVP ' Trypsin Digestion Lys Blocked Tryptic Peptides Sephadex G-25 Voids Sephadex G-50 Voids Sephadex G-75 Voids Figure 3. Fractionation Scheme of the Preparation and Purification of the Lysine Blocked Tryptic Peptides of GVP. l7 .co_um_>o_mz .: oenmwu I8 .: mcam_u mozama 8:635 222: o __ © \ I \ o I :0 I z + O __ Add In x I xw .. :0 o 8:23.22 19 2.6 N-Bromosuccinimide Peptides. The procedure used is outlined in Figure 5. One hundred mg of Fraction E was flash evaporated to dryness, resuspended in 88% formic acid, and stirred for 30 minutes. The acid denatured protein was again dried as above and then resuspended by cavitation in 27 ml of 50% glacial acetic acid. To this suspension was added 675 pM N-Bromosuccin- imide (NBS) in 27 ml of l00% glacial acetic acid (68). After stir- ring at room temperature overnight, the reaction mixture was flash evaporated and the entire NBS procedure was repeated with a reac- tion time of only A hours. It was then flash evaporated and applied to a column (2.5 x I00 cm) of Sephadex G-25 in 0.l M acetic acid, pH 2.87. The eluant profile was determined at 25A nm. NBS elutes with the monomer volume, Vm. The soluble NBS peptides were divided into three separate fractions: NI, N2, and N3; these were rechromatographed over Sephadex G-25 in 25% formic acid and analyzed for protein at 25A nm. The insoluble fraction NPT was eluted from the top 2 cm of the Sephadex G-25 column by successively washing the beads on a scintered glass filter (M) with both 88% and 25% formic acid. The filtrate which appeared flocculent was flash evaporated and again treated with NBS (as above) for A hours. The NPT peptide was effectively separated from the soluble peptides on Sephadex G-25, as before, eluted from the beads, dried and reacted with NBS for a third time. 20 Intact GVP (Fraction E) 88% Formic Acid Mix l/2 hour Acid Denatured GVP Dry (Flash Evaporation) Dry Acid Denatured GVP *Cavitation in Glacial Acetic Acid and N-Bromosuccinimide Mix Overnight - Dry NBS Peptides l Dry {Sephadex G-25 0.l M Acetic Acid I ; SolublelPeptides Insoluble NPT [T I Repeat from NI N2 N3 * above (2X) Sephadex G-25 25% Formlc Acid Ala-Val-(Val)-(Val)-Leu-Val-(Val)- N3 Ile-(lle)-Leu-(Leu)-Ala-(Leu)- FT—TT—TT I Nla N3 N2a N3 Val-IIe-(25 residues) +NZa Sephadex 6 - 25 0.l M Acetic Acid I I I NIA N2A NX Ala-Glu-AIa-Val-Gly-Leu-Thr-Gly-(Ser)-Ala-(Pro)-(Val)-Ala-Ala Figure 5. Fractionation Scheme of the Preparation and Purification of the N-Bromosuccinimide (NBS) Peptides of GVP. 2l 2.7 Digestion of the NPT Peptide by Thermolysin. The incubation mixture consisted of l00 nm of NPT peptide, IOO nm of norleucine as an internal standard, and ICC pl of Tris-HCI buffer 0.05 M, pH 8.0, containing 0.002 M CaCl Incubation was 2. at A00 C after addition of 0.05 mg thermolysin (Calbiochem). Aliquots were removed at 0, l, 3, 5, and I2 hours and the reac- tion stopped by the addition of glacial acetic acid to a pH below 2. All aliquots and standard amino acids were simultaneously subjected to electrophoresis at pH l.9 in a formic acid-acetic acid buffer system (69) in order to measure the extent of diges- tion. Subsequently a large sample of NPT (3 pmoles) was digested with equivalent increases in reagents fiarlO hours. The reaction mixture was chromatographed on a column of Sephadex G-25 in 25% formic acid. The insoluble NPT was removed from the column as described in the preparation of the NBS peptides. The soluble fraction produced an eluant profile which was analyzed with ninhydrin at 570 nm after NaOH hydrolysis (70). PolyprOpylene tubes were used as glass releases silicate which interfered with the analysis. To l-2% of the volume of each frac- tion,ordinarily corresponding to l5-30 pl, was added 200 pl of 5 N NaOH. All samples were autoclaved at l2loC for l hour, and then neutralized with 200 pl of 5 N HCl, and then brought to l ml with pH 6.l citrate buffer 0.0A M Ninhydrin analysis at 570 nm followed. 22 2.8 Partial Acid Hydrolysis of the GVP and Peptides. Partial cleavage of fraction E of the GVP was attempted with 0.03 N HCI at l05o C for A8 hours in a nitrogen atmosphere (7l). The rate of the reaction was followed by monitoring the release of free aspar- tic acid. The same hydrolysis procedure and assay method were used on peptide NPT and the hydrolysate chromatographed over Sepha- dex G-25 in 0.I M acetic acid. Fractions were monitored for protein at 25A nm. To prepare fragments of the TlP2 peptide, 2 pmoles of the peptide were hydrolysed with 200 pl of 6 N HCl at 105° c for 5 minutes. The hydrolysate was rapidly dried under a stream of nitrogen at 250 C and electrOphoresed at pH l.9. Peptides were eluted with 0.I N acetic acid, dried, and stored in 25% formic acid at -200 C. 2,9 Recovery of Peptides (Balance Sheet). The percentage of recovery of peptides in my work is based on the amount of lysine and arginine applied to the column vs. the amount recovered in the various fractions. Aliquots from the pooled fractions were assayed for lysine and arginine on a Technicon Amino Acid Analyzer. 2.l0 Purity of Peptides. High voltage electrOphoresis at pH l.9 and/or pH 6.5 (69) was used in conjunction with the cyanoethylation procedure of Fletcher (72) (Figure 6) and amino acid analysis in order to determine the homogeneity of the purified peptides. Both proced- ures were necessary as often very large peptides and peptides of 23 3 .co_um_>£uoocm .m we:m_u .w ocam.u OI we .. «118/ I0\ .0... 2w /..2 oN I m + :oNo .. 3.5: © 0.8. o 3.88 2.55386 mesa I zmor~.~.._o./ \1 +_. Hznz...@ I....lo z/ + zmououorx N zmo i~.~zo.\ I 5.6.5.0085 25 very low yield could not be visualized by a ninhydrin assay after electrophoresis. Cyanoethylation was sensitive to a minimum of 5 nanomoles of peptide. 2.ll Sequence Analysis of Peptides by Carboxypeptidase C. Certain peptides were sequenced from their C-termini using carboxy- peptidase C (Henley and Co.) using the method of Tschesche and Kupper (53). Norleucine was the internal standard. The reaction was stopped by adding 0.I N HCl to pH 2 and then the mixture was heated to 800 C for l minute. The samples were then centrifuged to remove the enzyme. The supernatant liquid was removed and analyzed on a Technicon Amino Acid Analyzer. 2.l2 Amino Acid Analysis. Samples were hydrolyzed in evacuated l ml microvials (Precision Sampling) containing 200 pl of "6 N” (constant boiling) HCl at IOSO C for l8 hours. The solution was then brought to dryness at 500 C under a stream of nitrogen and the residue resuspended in citrate buffer pH 2.875 according to Moore.a£_al. (73). Amino acid analyses were performed on an auto- mated Technicon Amino Acid Analyzer by the accelerated method using Chromobeads C-2 (7A). The chromatograms were integrated by an Autolab System IV Computing Integrator. Based on the retention times relative to the internal standard, norleucine, each amino acid was identified and the corresponding correction factors were applied automatically for the individual color reaction of each amino acid with ninhydrin. 2.l3 Automated Sequencing Methods. Peptides and proteins were sequenced on a Beckman Sequencer Model 890C, except where otherwise 26 noted. The various methods used to sequence the intact gas vacuole protein via automated Edman degradation (Figure 7) are given in Table I. All peptides were sequenced using the dimethylalIyIamine- trifluoroacetic acid buffer in a pyridine-water mixture (DMAA) with peptide program #02I572 as diagramatically represented in Figure 8. When possible the peptide carboxyl terminal modification method of Braunitzer (59) (Figure 9) employing sulphophenyl isothiocyanate (SPITC) was used. All other peptides were sequenced using my improved version of the method of Fosterla£_al. (60) (Figure ID). This method was devel- Oped as indicated in Table 2 by using a synthetic tripeptide (DL-Leu- Gly-Gly). Solution A: 300 nM peptide in 600 pl of water, adjusted to pH A.0. Solution B: 5 mg EDC (N-ethyl, N'-(-3-dimethylaminOpropyl carbodiimide) HCI and 8 mg ANS (2-amino-l, 5 naphthalene disulfonic acid) in l ml water, adjusted to pH A.0. Solution C: 5 mg EDC in l ml water, adjusted to pH A.0. Add to Solution A 20 pl of Solution B. Mix for 3 hours, then add 20 pl of Solution C. Mix for l hour or longer at room temperature. Use 20 pl of Solution C repeatedly for poorly solubilized peptides. At the end of the reaction, add 50 pl glacial acetic acid. Apply to reaction cup on sequencer in amounts of less than A00 pl. Dry via the Beckman sequencer subroutine #02772. 27 Table l Methods and Results of Sequence Analyses of Gas Vacuole Protein (Fraction E) Treatment Acid Heat SPITC DMAA Qpadrol Sequence + - denat. denat. #05077l #02672 Results + - - - + Ala-Val-Glu Gummy + - + - + residues + - + + - Ala-Val + + + + - Ala-Val + + - + - Ala-Val + + + - + Ala-Val + + - - + Ala-Val 28 .mpp.uaoa>.om mo m.m>_mc< cocoaqom Lem co.umumemoo cmEvm use .5 ocsm.u 29 eczeeeaeéaeed :2 . o. 28 /...\ ,e A o 88:00 .m._ 032mm. «:0 SEE «Baud 055.325 05.8 mioiz I _._. \ / 3...... sorormxz + 08 z a. /...\ .n. m IVZ 06 i m 8a m > .8 a I I0/ Io.m In I 2. 1 Sec 8.58.. a; _ e 3:... . i \ IQIoIZIo .112 All QIoIZIoIoIz + may; m/o\ 44.20 mum. m .m._ x: mu uz 5:36.000 58.0w 30 PEPTIDE PlTC/HEPTANE Then N2 dry PITC + PEPTIDE Dimethyl allylamine buffer for COUPLING REACTION (900 seconds) PTC - PEPTIDE Benzene wash (Additional 300 seconds coupling) PTC - PEPTIDE EXCESS REAGENTS TO WASTE N2 DRY DRY PTC - PEPTIDE I Benzene wa§h I WASHED PTC - PEPTIDE BENZENE WASH T0 FRACTION COLLECTOR N2 dry then Hepta fluorobutyric acid CLEAVAGE PEPTIDE (n_]) + thiazolinone - amino acid TO FRACTION COLLECTOR L Butyl chloride wash __a r 1 PEPTIDE (n-l) thiazolinone - amino acid N.HCl+Et SH l0 min 80° PTH - amino acid Figure 8. Operational Scheme for Peptide Program #02I572 for Beckman Sequencer Model 890 C. 3] .Auh_¢ my mum cm>oo_ .LHOm _ _>c ocaoca_:m mc .m: :0 .umo_ .m.voz m Bede a .m ocnm.m 32 .m mezm.u eczema 3.58 3.2.3.: 20550.58:323:3 m3 _oc.E§Io Av AV cum cum .0. - , .. I» o I u .. .. r I momI°IzIWIZI£~IQMV ®<<2mom z Wrmzfl£uxgxo ® .. ._. _ 8:828: deeded .6 8522 52.55 33 .n_u< u_co._:m mcm_m;u;amz oc_e< ucm mv.s__vongmu >3 co_umu.._noz mu.uamm .o. m.:m.m 31+ .0. m.:m.. 3.32. 8.23: v.8 25:3... 2.222%: 0.. LEE... IN mom one + . IoI©A| u n Ia I \ n ‘ Iz . ._._ mom «1182 €5.12; 00 lz 0.0 + mom a: 8.5.53.8 232.3...52336 In. .. . z .......?z 8.32. .o z a» an I u I I n z I «2 ‘oI@AI . :22 . :8 z o z «:o zo+xxo @ o 06:... o 828.382 8:8. ,6 8522 .28“. 35 .Aomv ._m we .mumou mo venues 0:» ._: ooo. ucm oow cmmzuon mm: 0.:ux.E co.uumu. osu mo oE:_o> och .m.; 3 mm: we.u co.uumo. .mHOu ugh u .umz z _.o J o m. o m. m. m I J m m._ m m. m - a o m; o m. m. m m a J o m._ o m. m. m n. - a o o. o m; n. q u u m o o. o m. m. m 0 I o o o a a o o o m m QN m I J o o o o m. _ m.: x: m.c z: z: 2 .2... 5:33 8.3.03 .353. m0 oE.h acaoe< .0 we.» u::0E< >_on>_un:o. ucOEumOLF mco.u.uv< In uom mz< ou.uaom acoEumo.h mgauouo.m co.umu.m.uoz ov.ua0¢ vo>o.ae_ N 2.5 36 2.1# Identification of Amino Acid Derivatives from Seguence Analysis. All PTH amino acids were formed from the thiazolinone as follows (75): Add 0.2 ml of 1.0 N HCl containing 1 pl/ml ethanethiol to dried fractions. Blow nitrogen over solution for 30 seconds and seal with silicone rubber stoppers. Mix by cavitation. Heat to 800 C for 10 minutes. Cool quickly. Extract PTH with 2 volumes each of 0.8 ml of ethyl acetate. With the exception of histidine and arginine, which is destroyed by gas-liquid chromatography (GLC), all PTH amino acids were analyzed using GLC employing the SPOV column matrix of Pisanoig£_gl. (76). The temperature program used was: Initial temperature: 1700 C. Delay: 2 minutes. Rate: 6o/minute. Final temperature: 2800 C, hold at this temperature for 12 min. The on-column method of PTH-TMS derivatization was performed with N,0-bis-trimethy1-silyl-acetamide (BSA) in order to identify the more difficult PTH amino acids. However, due to difficulties concerning very low yields in vacuole peptides, Asn, Gln, Ser, and Thr were identified after hydriodic acid hydrolysis (1200 C, 18 hours, N2 atmosphere) according to Smithies'g£_gl. (77) and Inglis‘g£_gl. (78). The hydrolysates were then analyzed for amino acids using the previously described citrate buffer system. RESULTS 3.1 Purity of Gas Vacuole Membranes. The purification of the gas vacuoles (2.2) was monitored by growing cells in the presence of 35 S and by assaying the cell free fractions (Figure l) for their specific activity (2.3) based on the fact that gas vacuoles do not have any sulfur containing amino acids. Table 3 summarizes the results and shows that Fraction E, which was used for peptide preparations, has a relative purity of 99%. 3.2 Amino Acid Composition of Gas Vacuole Protein. The amino acid composition of gas vacuole protein (GVP) has been reported before (39). As an improved analysis technique became available, I repeated these determinations, the results of which are summar- ized in Table A. The relative amount of some amino acids changed slightly; noteworthy are alanine, valine, isoleucine, and leucine and the occurrence of proline. 3.3 Sequence Analysis of GVP. A common procedure in sequence analysis using the Edman degradation is to obtain an amino terminal sequence of the intact protein: sixty consecutive residues were obtained from apomyoglobin D of the humpback whale (75). I tried such an approach with GVP as shown in Table 1: however, a maximum of 3 residues A1a-Va1-Glu-...was obtained. The possibility of protein loss from the reaction cup due to washout, was ruled out by amino acid analysis of the cup contents. 37 38 Table 3 Purity of Cell Free Fractions Containing Gas Vacuole Membranes Fraction cpm per Protein Specific Activity Relative assayed (Fig. 1) 0.2m1a ug/0.2mla cpm/ ug Contamination A 225092 550 8 317424 430 738 100% C 90865 150 605 81% D 5761 60 96 13% E #16 50 8 1% F 77 35 2 0.2% Average of two determinations. Il'r ‘ 39 Table 4 Amino Acid Composition of GVP Ami no Acid mole%a mole%b mole%c mole%d Aspartic Acid 6.3 6.h 5.6 11.6 Threonine 4.5 5.2 4.9 6.5 Serine 9.3 9.3 9.9 8.1 Glutamic Acid 9.2 11.6 12.4 8.6 Glycine 9.5 3.3 4.3 9.4 Alanine 15.2 18.6 15.9 14.7 Valine 15.0 11.6 12.6 8.8 lsoleucine 10.3 7.3 10.1 9.1 Leucine 9.3 11.2 10.0 0.9 Tyrosine 2.4 3.3 2.8 2.3 Phenylalanine 0.4 0.0 0.6 2.5 Lysine h.2 5.4 h.7 h.7 Histidine Trace Trace 0 2.5 Arginine “.0 4.8 4.2 5.5 TrytOphan n.d. 1.5 0.7 n.d. Methionine 0 O 0 O Cysteine 0 0 0 0 Proline 5.6 Trace 1.h h.2 Current data. These values are corrected for the decomposition of Asp, Thr, and Ser and the release of Val and lle. Data from Jones and Jost (39). Composition of gas vacuoles from Anabaena flos-aquae (A7). Composition of gas vacuoles from Halobacterium halobium (48)- 3.h Estimation of the Number of Tryptic Peptides. Jones (AS) fingerprinted the tryptic peptides of the GVP with ninhydrin. How- ever, incomplete trypsinization resulted in a variation of the pattern (See 3.6). I approached the problem by dansylating amino terminal amino acids of the tryptic peptides. After hydrolysis of the peptides the dansylated amino acids were separated and visu- alized. Figure 11 shows a minimum of 9 separate spots, two of which do not correspond to dansylated amino acids. Thus, a minimum of 7 peptides is indicated. 3.5 Separation of the Tryptic Peptides. Figure 12 represents a typical profile of tryptic peptides separated on Sephadex G-25. The profile as determined by ninhydrin analysis was always very reproducible, whereas the separation of the S , S , S peaks as 1 2 3 monitored in the ultraviolet, was inconsistent. Thus, the ninhy- drin fractions, T1 and T2 were used. Further separation of the fractions T1 and T2 (Figure A) is shown in Figures 13 and lh. The first peak, TlPl, can be separated on Sephadex G-25 in 25% formic acid into 3 major fractions (Figure 15). TlPla is a mixture of an arginyl peptide, a lysyl peptide, and some T1P1b peptide, as determined by amino acid analysis (Table 5) and by carboxypeptidase C digestion (2.11). Peak T1P1b is a large frag- ment of about #5 residues (Table 5). This large fragment results from incomplete trypsinization. The third major peak in Figure 15 contains some overlap from TlPlb, NH#+, and NaCl. Figure 11. Al Minimum Estimation of the Number of Tryptic Peptides. Tryptic peptides (2.A) were dansyiated and hydrolysed according to Gray (65) and separated electrophoretically at pH 1.9, SkV, 135 minutes. Standard dansyl (DNS) amino acids were applied as indicated. #2 o 1 b DNS Lys. 0 h 70 - ' ° DNSArgo . f DNS His. 0 O DNSAIO. DNSTy'O : g D 39' DNSVaIO so mama: ”Mat,“ .. E DNSPM sp'ons Pro. ' a o DNSTryO 30 - pH l.9 _. 135 min. 5 KV IO - Origin J l 1 J I (Standard DNS Amino Acids) gas Vacuoles Figure 11. #3 Figure 12. Separation of the Tryptic Peptides on Sephadex G-25 in 25% Formic Acid. T1, T2: fractions identified by ninhydrin and pooled for further analysis. S]. 52, S3: fractions identified by U. V. absorption at 280 nm. S Tryptic Pe lide on Sephagex IIo ' 50 90 Fractions (1.5 ml / tube) Figure 12. 45 Figure 13. Separation of Fractions T1 and T2AVI on SP Sephadex C-25-120. Elution was with a biphasic NaCi gradient (0.05 M - 0.7 M) in pH 2.3 formic acid and water (2.A). ISee Figure 1h. “ n + T2AV on SP Sephadex C-25-I20 in pH 2.3 formic acid .15- “PI 0.05-0.7M NaCI 1192 o O ‘2‘ .I0- 0. o TIP3 .05- gl 50 I00 Fractions ( 2m! I lube) Figure 13. M ‘NaCl A7 mm: co.u:_u .A:.~V aco.vmcm :a 6.0m u.uoomuac.u.c>a m zu.3 .mu< xmc.E< :0 NH co.uomcu o.ua>.h mo co.um.maom .J. o.:m.u .a. mtsm.. .23: .2 99.5.32... 1 100.0 0.0 9.0 on. o: om 0... on on o. a d LT d o.~ 0.9. 0.2. 0.9. 3236 I 3.2.5. .74 .8564 :o N... 0.85... Fun-nu-IIIIIIIIIIIIIIIII MI 01.9 .G'O 49 Figure 15. Elution Profile of Tryptic Fragment TlPl on Sephadex G-25 in 25% Formic Acid. Fractions TlPla, TlPlb are designated in the figure by the bars. 0. Do 280 .10 .05 50 TIPI on Sephadex -25 in 25‘%. Formic Acid TlPla TIPIb i--4F-—-4 Vo Vm l—-lr: 1 1 1 IL ‘1‘ 1 50 100 Fractions (l.5ml / tube) Figure 15 of the Purified Tryptic Peptidesa 51 Table 5 Amino Acid Composition Amino Peptides Acid TlPla TlPlb TlP2 T2A2 T2A3 T2Au 71928 Asp 3.5 3.2 1.0 0.3 Thr 2.5* 2.8 1.0 Ser 5.9* 5.8 1.2 0.2 Glu 5.5 5.8 2.1 0.2 0.7 Pro 3.0 2.8 0.9 0.8 Gly 3.0 3.6 1.0 0.9:: Ala 7.3¢ 8.9c h.9¢ l.9 0.7¢ l.8¢* Val 3.9 h.h 2.0 0.3 0.7 0.9 lie 1.3* 1.1 1.1 0.3* Leu 3.0 3.3 1.1 0.3 Tyr 1.2* l.h* 0.8* Phe 0.6 Lys 0.9¢ l.0¢ 1.0¢ Arg l.0¢ l.0¢ 1.0¢ Total h2.6 hh.l 15 6.6 h.0 1.0 3.5 Values are expressed as residues per peptide. This peptide fraction is heterogeneous. This peptide is a fragment of TlP2. Amino terminal amino acids. Carboxyl terminal amino acids. The lysyl peptide was selectively modified and subsequently sequenced (3.6). 52 If peak TlPl is separated over Sephadex G-25 in 0.1 M acetic acid (Figure 16), the peptide T1P1b elutes with the void as expected for a #5 residue fragment. The drastic difference in separation of this peptide on the same gel matrix with different solvents sug- gests that the highly protic 25% formic acid may be inducing ion- exchange qualities in Sephadex G-25. The shouldered peak TlP2 contains the carboxyl terminus, peptide TlP2 (Table 5), of the GVP and a small amount of contamin- ating peptides, perhaps some overlap from fraction T2. TlP3 is an agglomeration of several peptides which partially originate from T2. The large T2Al peak as seen in Figure 1%, consists mainly of ammonia and at least one unidentified peptide as based on the pres- ence of both arginine and lysine. The T2A2 shoulder contains pri- marily a unique arginine peptide (Table 5) with some overlap from T2Al and T2A3. T2A3 is a mixture of two lysyl tetrapeptides, T2A3a and T2A3b, which consistently occur in unequal proportions (Table 5). Attempts to separate T2A3a from T2A3b by electrophoresis at pH 6.5 failed because the two peptides have the same size and charge. Fraction T2Ah is free lysine (Table 5). The T2AV fraction elutes with the void volume of the Aminex A-5 column and was combined with fraction Tl. T2AV contains the overlap of T1 into T2. The recovery during separation of the tryptic peptides (2.9) is given in Table 6. More arginyl peptides than lysyl peptides are 53 Figure 16. Eiution Profile of Tryptic Fragment TlPl on Sephadex G-25 in 0.1 M Acetic Acid. 0. D. 280 TIPI on T Sephadex G-25 in 0.1 M Acetic Acid .15 *' TlPlb .IOr .05 ’ 5h 50 100 Fractions (2 ml/tubcl Figure 16. 55 Table 6 Recovery of Tryptic Peptides Fractionation u Moles % Yield Stepa Lys Arg Lys Arg Fraction E 33.0 31.0 100 100 Acid Precip. 33.0 31.0 100 100 GVP Tl 8.h 6.h T2 22.5 1#.h Column residue .7 1.1 Total 31.6 21.9 96 70 T2AV 1.8 2.2 T2Al, T2A2 8.5 1.7 T2A3, T2Ah Tatal 1003 309 3' '2 TlPl 1.5 1.8 TlP2, TlP3 1.7 .5 Total 3.2 2.3 9.8 7.5 See Figure 2. 56 lost. Losses are due primarily to irreversible binding of various peptides to both Aminex and Sephadex matrices. Sephades G-25 e.g. retains 30%.of the arginyl peptides of the fractions T1 and T2 (Figure 2), whereas virtually no lysyl peptides are lost. I also observed that a new batch of SP Sephadex will irreversibly bind more than 95% of the peptides applied to the column. With each succeeding application, this loss decreases. This also seems to be the case with Aminex A-5 and Aminex AGSOW-XZ. Complete amino acid analyses of the tryptic peptides are presented in Table 5. The amino and carboxyl termini were identi- fied by cyanoethylation (72), and when necessary, by carboxypep- tidase C (53). Lysine and argine, when present, are considered in Table 5 as carboxyl termini on the basis of the specificity of trypsin. The peptide TlPIa is heterogeneous, but the peptide modi- fication method of Braunitzer (59) allows selective sequencing of the lysine containing peptide. 3.6 Sequence Analysis of the Tryptic Peptides. This was done automatically by carboxypeptidase C and/or by automated Edman degradation. Carboxypeptidase C is an exopeptidase which removes all protein amino acids including lysine, arginine, and proline from the carboxyl terminus of a polypeptide. Peptides must be rath- er pure to obtain a reliable sequence. The carboxyl termini of tryptic peptides T2A2, T2A3, T2Ah are either arginine or lysine (Table 5). 57 The carboxyl terminal peptide T1P2 was treated with carboxypeptidase C. The data in Figure 17 show 2 alanines as the first amino acids liberated. After fragmentation of peptide TlP2 by partial acid hydrolysis (2.8), only one homogeneous peptide fragment, TIP2a, was isolated. The amino and carboxyl termini of the peptide are alanine (Table 5). Treatment with carboxypeptidase C suggested the sequence: Ala-Pro- Val-Ala-CODH but the yields were low and the amount of peptide limiting. Fragment TlP2a is considered to be adjacent to the car- boxyl end of the peptide TlP2 based on the observation that partial acid hydrolysis would result in cleavage at seryl, or aspartyl resi- dues and that Ala-Ala-COOH is the carboxyl terminal sequence of pep- tide TlP2. Thus, a sequence of Ala-(Pro, Va1)-Ala-CO0H could appear nowhere else in the peptide TlP2 (Table 7). Washout of the peptide TlP2 from the reaction cup prevented completion of the sequence. Chemical modification of the peptide increased the yields but did not allow a complete sequence determination either. I The asterisked amino acids, Ser*, Glu* (Figure 17) and Ser*, Leu* (Figure 18), probably originate from the digested peptide which had 1 or 2 alanines removed by the enzyme and thus, are not part of the carboxyl sequence of either the peptides TlP2 or TlPlb (Table 7). The Technicon Amino Acid Analyzer also separates and displays small peptide fragments which may be eluted with the same relative retention time as amino acids. Consequently, the use of carboxypeptidase C is limited as a sequence tool since such spurious peaks appeared in all assays. 58 Figure 17. Release of Amino Acids from Peptide T1P2 by Carboxypep- tidase C. Deduced Sequence: Ala-Ala. See Footnote 1 (3.6) for explanation of Ser* and Glu*. moles x ID"9 of released amino acids 30 B9 (I N O 51' 5 59 on TIPZ _ Carboxypeptidase C 0 Ala ll 0 Ser‘ 1."'---'p [.7 . G'“ * 30 60 9 120 150 Minutes Figure 17. 60000.0. 300 0...- .2220... 0...» we «:3... 0... 00:06.9. 3 000.. .. C0800. .0 09000020.. 00. 3.0 0003 8.00028 0..“ 3 0.02.3. 70.30... :0 00 003.0 3 058—3.. 2: "0.303.... 9.16:2 0... :o 00.0.. ._ 30; 00003020.. 2....0 .. .2032} 0.00 3.9.0.. 3 000.030 002050.. 00 8.00020 9 300.0%00 no .u 003.000.580.00 .2 .E 5.3 3.320%. 2. .00.»..300800 "ESE. 05 we 30 3 8.0.000 c. 00... 8.0002390... 03.00: 00 8500 05 000060.000 08.0 8.3023 0E. x .3 2n. 0 ON a 0 z .8. 03 30¢ :03 0— _ 35... 0 g 3 0 60 a... a... 2.. .2 .2.» s a. 2 am an a u u a. .. B2 .2 22. 2.... ... .3. ... .8 «<3 3 R «m R an 0 .9. u. u. u. .. .2 .2 2.... .9... 22. 3... .20....» .3. .33 ..> .2 3. .2 .3 a...» a : a... a... m a n... o. a... n. o. «a a... «d u .. .. .. a .. .. .2 .25 i» 92.33.... m: 2.. 2.. .2 .2 .8 a... :n a... to .2. .2 30 .2 .a. .2..— «d m... a... 0.2 n.» or : u. a... a. 3 8 .32... z .. .. .. z .. :2 3...... .3 1.3.8.3313... 22. 7... .2. 2. .2. a... :u .2. .2 a... .2 .3... .82....» 0:: ”an savanucnnNNN—«oug a. 2 o- m— .1 n— N. Z 2 m o s o m a n a — .59 I9.. .0308. 0.0%» 0.3 00 00020.2 02.0.63 5 0.00... 61 Figure 18. Release of Amino Acids from Peptide TlPlb by Carboxy- peptidase C. Deduced Sequence: Ala-Ala. See Footnote 1 (3.6) for explanation of Leu* and Ser*. moles xlO” of released amino acids 62 Carboxypeptidase C 22 ’ on TlPlb ‘- ‘0 Ala N O - O I I N 6 .Ser" Hours Figure 18. 63 Peptide TlPlb was only partially sequenced by automated Edman degradation (Table 7) despite modification of the carboxyl terminus (2.13). Although the peptide contains arginine, carboxypeptidase C releases 2 alanines (Figure 18) from peptide T1P1b. Thus the amino terminal portion of peptide TlPlb must be linked via arginine to the TlP2 peptide. The peptide TlP2b comprises the carboxyl terminus of GVP. Peptide T2A2 was only partially sequenced by the automatic procedure due to peptide washout. Good information regarding the carboxyl terminus of the peptide T2A2 was obtained (Figure 19). The sequence is: (Asp)-(A1a)-Ala-Arg-CO0H. There is no serine in peptide T2A2 and glycine is the amino terminus as determined by cyanoethylation (2.10, Table 5). Thus, Ser* and Gly* are not a portion of the carboxyl terminal sequence of T2A2. The apparent sequence of peptide T2A2 is given in Table 7. Peptides T2A3a and T2A3b were sequenced by an unconventional method based on the presence of only 2 peptides occurring in very unequal amounts. A mixture of 2 peptides from tr0poe1astin was analyzed by a similar approach by Foster g£_gl. (60). The yield ratio must be consistently unequal for the 2 peptides. The concen- tration of amino termini of peptides T2A3a and T2A3b differ propor- tionally i.e. the amount of isoleucine was less than that of alanine (Table 5). Sequencing of peptides T2A3a and T2A3b modified with SPITC proceeded as usual except that each cycle produced 2 residues in amounts as expected for the relative concentration of the 6h Figure 19. Release of Amino Acids from Peptide T2A2 by Carboxypep- tidase C. Deduced Sequence: (Asp)-(A1a)-Ala-Arg. See text (3.6) for explanation of Ser* and Gly*. males x 10" of released amino acids 65 :4- r3 U 6 lb Carboxypeptidase C onT2A2 0 Ala Arg .Ser"E §L 60 IZO ISO 240 Minutes Figure 19. 66 peptides. As seen in Table 7, the third residues were in equal yield so assignment of these amino acids to their respective pep- tides was based on other data. it was already known from amino acid analysis that the concentration of each of the amino acids (alanine, valine, glutamic acid, lysine) in peptide T2A3a was greater than the concentration of the amino acids (isoleucine, leucine, aspartic acid, lysine) in peptide T2A3b as shown in Table 5 so the conclusion that glutamic acid belonged to T2A3a and aspartic acid to T2A3b was made. Lysine was placed at the car- boxyl terminus of each peptide on the basis of the specificity of trypsin. A partial sequence of fraction TlPla was also obtained by an unconventional method. The amino acid composition (Table 5) shows that peptide T1Pla contains a mixture of several peptides having different carboxyl and amino termini. SPITC reacts with only the free amino groups in peptides (59). As all unmodified GVP peptides readily wash out of the sequencer reaction cup within 2-3 cycles, the peptides terminating in lysine should remain. Based on the appearance of 9 consecutive unique residues (Table 7), the possibil- ity of several peptides terminating with lysine was also discounted. All other T1 and T2 peptides were not obtained in either sufficient yield or purity to warrant sequence analysis. 67 3.7 Maleylated Tryptic Peptides. Overlapping peptide sequences are needed, in order to align peptide fragments. Maleylation, partial acid hydrolysis, and N-Bromosuccinimide hydrolysis of GVP were tried. Blocking theez-amino group of lysine often enhances solubility of proteins and prevents cleavage of lysine by trypsin. Therefore, I tried to obtain lysine-blocked tryptic peptides. The advantage of maleylation (67) is its reversibility below pH 6.5. The trypsin- ized, maleylated peptides were chromatographed and demaleylated over Sephadex G-25 with either 25% formic acid or 0.1 M acetic acid pH 2.87 (25). The elution profile of fraction T1, T2 is the same as in Figure 12. These fractions were then pooled and chromato- graphed (Figure 3). Elution profiles were identical to those for tryptic peptides (Figures 13 and 1h). The amino acid composition of the peptides of T2A3 does not differ from that of unmaleylated protein. Apparently no maleylation of the lysines of the unboiled GVP occurred. 0n the other hand, heat denatured, maleylated GVP was found to have all of the c-amino groups blocked (2.5). By cyanoethylating the protein before and after maleylation, the per- centage of lysines blocked was calculated. This maleylation pro- cedure was found to be 100% effective. After the tryptic treatment (Figure 3), the digest was chromatographed over Sephadex (2.5). The material voided all three gels. Since the GVP is very hydro- phobic and poorly soluble, it is possible that either untrypsinized aggregates of the GVP form or that a few peptides have a tendency to associate and precipitate as the eluant at the void is rather turbid. 68 3.8 Dilute Acid Hydrolysis of GVP. Hydrolysis of protein with dilute acid (2.8) can produce peptides by cleavage and release of aspartyl residues (71). However, no free aspartate was found in the hydrolysate mixture with GVP. 3.9 N-Bromosuccinimide (NBS) Peptides. Tyrosyl and tryptophanyl residues in protein can be attacked and degraded by NBS (68). To achieve maximum cleavage of the GVP (2.6), 8 moles of NBS per mole of (tryptOphan + tyrosine) were used. The extent of reaction of GVP with NBS was monitored by the appearance of peptide NPT (2.6) as measured by amino acid analysis. NBS was applied 3 times. The NBS hydrolysate of GVP was chromatographed on Sephadex G-25 in 0.1 M acetic acid, pH 2.87 (Figure 20) and the three major frac- tions N1, N2, and N3 were separately fractionated on Sephadex G-25 in 25% formic acid. Figure 21 is a composite of these elution patterns. Fraction N1 separates into fractions Nla and N3. Frac- tion N2 separates into fractions N2a and N3. N3 elutes with the salt. The amino acid analyses of these different fractions are given in Table 8. Peptide NlA has alanine as its amino and carboxyl termini. The elution profile for Nla and N2a is given in Figure 22. As shown in the figure, more than 70% of the pooled fractions contain the single peptide NIA. Recoveries (2.9) of NBS peptides during separation and purification are given in Table 9. Cyanoethylation of the amino terminus of peptide NPT (Table 8) reveals that there are about 40 amino acid residues in the peptide. 69 Figure 20. Separation of the N-Bromosuccinimide (NBS) Peptides on Sephadex G-25 in 0.1 M Acetic Acid. 70 T 7"] NBS Peptides on Sephadex 6-25 in _ 0.I N Acetic Acid 0"5 pH 2.87 0.05 . |—-I-N 2—l—N 3——l 0.0 I Vrn I I‘ I I? I I I I 1 so 70 so no 130 150 Fractionsi2.5 ml / tube) Figure 20. 71 Figure 21. Composite Elution Profiles of the NBS Fractions N1, N2, and N3 on Sephadex G-25 in 25% Formic Acid. (- - - -) is N1; (....) is N2; ( ) is N3. Fractions are designated as Nla, N2a, and N3. .|.8_x. .8 do 72 ‘ 70 Fractions (1.5 ml / tube) O 5 I .l. O O . q. \q u q» 141 u . u u - Q \ \\0L 3 ““‘O”OOVA' n N \|||.nu ........ 0.I...” .11.. 3 . 3.000 ' I 0‘2““ N OOOOOOO’ON""'- ”. ..... . 000’ .060":v a 00 \~ ........ m 2.. iii px 9! N00 .I“‘ earn Mntliiial P a o N "" 0000 3”? to:llt.l.:.. mmw.“w “m" al.0JuIlmW - p h 0 5 0. 5 l. m m o. o 0 0 110 Figure 21. 50 73 Table 8 Amino Acid Composition of the NBS Peptidesa Amino Peptides Acid NIA N2AC N3c NPT Asp (0.5) 1.6 1.7 1.5b 3.0b Thr 1.2 1.3 1.0 1.0 2.0 Ser 1.1 1.9 1.3 2.A h.8 Glu 2.3 h.3 2.8 2.3 h.6 Pro 1.1 O.A 0.9 0.7 l.h Gly 1.1 1.9 1.5 1.1 2.2 Ala S.5¢ h.0 3.2 3.5* 7.0* Val 2.2 1.9 1.6 3.0 6.0 lle (0.2) 0.5 0.“ 1.7 3.h Leu 1.2 2.1 2.2 2.3 h.6 Tyr Phe 0.5 Lys (0.2) 0.6 1.8 0.9 1.8 Arg (0.1) 1.0 1.0 1.0 2.0 Total 16.7 21.5 19.9 22.0 #2.8 a Values are expressed as residues per peptide. b The first set of values is based on the presence of a unit arginine. The second set of values is based on an amino terminal analysis of the peptide by cyanoethylation. c These peptides may be heterogeneous. () Residues in parentheses are fractional residues thought to be impurities. Amino terminus. ¢ Carboxyl terminus. 711 t u m C C L I . mm — . N . C * .NN o.:m_m 75 .NN otam_. .832 .5 0.8 22.8... 030». ON. 0: OO— 00 00 Oh om e :5 o> .1... 422 squza 2... 2.8.. 2.3 mm... 3828 .5 32.122 _ J .06 no.0 “330:0 76 Table 9 Recovery of NBS Peptides Fractionation u Moles %Yie1d Stepa Lys Arg Lys Arg Fraction E 28.0 27.0 100 100 N1, N2, N3 6.4 5.0 23 19 Nla 1.1 .5 N2a .3 .5 3" "3'5 1.5 03 Total 2.9 1.3 10.5 5 NPT 6.0 7.0 21 25 See Figure 5. This yield is low as the large amount of the carboxyl terminal peptide has no arginine or lysine. 77 This estimate of peptide size is not completely reliable, as the large number of aliphatic amino acid residues are linked to one another so as to be poorly acid hydrolysed. A carboxyl terminal amino acid in peptide NPT was not found even after extensive treat- ment with carboxypeptidase C. This may be due to the extreme insolubility of peptide NPT. 3.10 Sequence Analysis of the NBS Peptides. Peptide NIA has as a carboxyl terminal sequence, Ala-Ala-COOH, when treated with carboxypeptidase C (Figure 23). Small peptide contaminants in the peptide NIA preparation do not permit further sequence analysis.2 The best, although not complete sequence analysis of NIA was accomplished through use of the automated sequencer on EDC-ANS modi- fied peptide. By comparing the sequence of peptides TlP2 and frag- ment TlP2a, a tentative sequence of NIA is proposed in Table 10. In order to improve the sequence analysis of peptide NPT, modification with SPITC and/or EDC-ANS was used (2.13). Although The asterisked amino acids, Leu* and Thr* (Figure 23) probably originate from either the undigested portion of peptide NIA or the contaminating peptides found in the NIA preparation based on the reasons given in Footnote 1 (3.6). Automated sequence analysis has placed leucine and threonine as the sixth and seventh residues in the peptide, so they could not be part of the carboxyl sequence of peptide NIA. The other amino acids shown, proline, valine, serine, and glycine (Figure 23) are also subject to the same criticism and therefore can not be positioned into the sequence. Small amounts of lysine and arginine were also detected. Their release is not shown as they do not occur in peptide NIA. 78 Figure 23. Release of Amino Acids from Peptide NIA by Carboxypep- tidase C. Deduced Sequence: (Val)-(Pro)-Ala-Ala. See Footnote 2 (3.10) for explanation of Leu*, Thr*, Gly* and Pro, Val, Ser. males x IO" of released amino acids 0 O NI 0 03 O (I O 3 0‘ O N O O 79 Carboxypeptidase C on NIA / Ala /. O I at Leu 0 Thr * 0 Pro / /0 Val /. G" 7‘ 2 4 6 8 IO 6.1.0.}: 3. e. .0330: so.» 0.529.004 gm... 3 I... €026.33 an: a. .. .2... 8. 8.22:2... .32.. .o 3...... .5 3. 80 n m.n .. a .6 v a .6. .u no u u .. .. .. .. .. .. .. .2 .2 .2... .9... 22...... a... .fi =... .8. 2.. .2 =2 .2 2:. n6 u m.n. an .— .. .. .. 8. 703.3... .0 .21.... 5 .E. A a. a... .3. 0: £3 .6 ad a.“ a.“ : : a... «.a .6 2 «A o. 8 9 an .22.; .. .. .. .. .. 3...... 2.3.... .o .38.. 5 2.: .2... .3... .2 .3... 3. 2... .: .2... 2.. i. .2... .2... 2.. .2 8.0.6.... r... 282......«222892222...n...:2. u . o n .. n ~ .. .22... .32... go 8.; 33:0; 8.. 0.3 we nag—0.! 09.0.60» 0. 030... 8| addition of 90 pmnles ANS and 5 or 6 treatments with 60 pmoles of EDC ‘was made over a period of 72 hours, the amino acid sequence of peptide NPT is still questionable (Table 10): overlap from numerous repeating, aliphatic residues makes the data difficult to interpret. There is in peptide NPT, for example, considerable overlap of residue number 5, leucine, into residue number 6, valine. Thus, such overlap makes the repeating sequences of resi- dues number 3 and h, and number 9 through ll questionable. A calculation of percentage yields of these residues does not allev- iate these ambiguities as the peptide is progressively lost with the organic washes. Hydriodic acid hydrolysis of the PTH-TMS derivatives of peptide NPT (2.3) gave only very low yields as the background washout was higher. This was also the case when the ethyl acetate fraction was used (2.lh). However, the data were very reproducible suggesting that an unusual sequence exists. Improved modification of the carboxyl end of the peptide was not feasible due to insolubility of the peptide. Thus, attempts were made to fragment peptide NPT by either trypsin (2.A), thermolysin (2.7), or partial acid hydrolysis (2.8). Peptide NPT contains both arginine and lysine. Figure 24 gives the elution profile observed after tryptic digestion. The recovery of the material separated is given in Table ll. Sequence analysis of peptide fractions TNPTa, TNPTb, TNPTc was prevented by mechanical difficulties and by the presence of several peptides. 82 Figure 2h. Elution Profile of the Tryptic Peptides TNPTa-e of NPT. Digestion with trypsin was 72 hours. Elution was with 25% formic acid on Sephadex G-25. 0-0254 Sephadex 6‘25 in 25 °/o Formic Acid 9 0.059- d a c b i—e: : + a 0.0l V0 Vrn 83 Trypsinized NPT on 50 60 70 BO 90 I00 "0 Fractions (l.5 ml/tubel Figure 24. 8h Table ii Recovery of the Tryptic Fragments of NPTa Fraction Io'9M Asp %Yield TNPTa 6h0 l6 TNPTb 450 i l TNPTc 530 13 TNPTd #20 IO TNPTe 300 7.5 NPT #200 100 Percent recovery is based on Asp present initially and in the subsequent fractions a - e. 85 Five hour incubations of peptide NPT with thermolysin produced 3 major fractions as observed after electrOphoresis. However, their yield was too low for sequence analysis. Hydrolysis for 10 hours produced a large number of short di- and tripeptides as indicated by the Sephadex G'ZS eiution profile (Figure 25). The overlap of the fractions and the small size of peptides precluded sequencing. The large peaks seen early in the elution profile (Figure 25) may be ammonia. Peptide NPT contains 3 aspartyl residues, thus, 3 or A peptides are expected after dilute acid hydrolysis. After 2h hours, more than 70%.of the aspartyl residues were released. Chromatography of the hydrolysate is illustrated in Figure 26. Major fractions resolved are: DNa, b, c, and d. The recovery of these peptide fragments is given in Table 12. Amino acid analyses (2.l2) of these fractions and NPT and fraction DNPT are given in Table l3. Cyanoethylation analyses (2.l0) of fraction DNb indicated the presence of more than one peptide as shown in Table l3. Free aspartate was the primary constituent of fraction DNc, representing over 37% of the initial aspartyl residues. DNd included overlap of free aspartate from fraction DNc and ammonia. Since fragment DNb contained lysine, selective modification using SPITC was used (2.13). Results are in Table l0. More exten- sive hydrolysis with either dilute or concentrated acid may permit additional sequencing. As the peptide NPT contains so many 86 .muuo xovmzaom :0 0.00 0.2.0. xmm cu.3 mm; co.u:.m .c.m>_oe.mzh >3 vmmmo_m¢ mov.uamm kmz mo co.um.maom .mN 0.3... 87 . .mN 0.3... .322... m... 22.8.... 2.... 253... $8 5 mm... .8235 km 2 no .20.. 52.35.05. _.O (pazfilmpfiu HODN) 0‘9 '0'0 88 Figure 26. Separation of the Peptides DNa-d Released by Dilute Acid Hydrolysisoof Peptide NPT. Hydrolysis conditions: 0.03 N HCl, l05 C., #8 hours. 89 0.34 ‘fiDilute Acid Hydrolysis 1’ of NPT on Sephadex G-25 0.I N Acetic Acid 0J5 ' pH 2.87 Is A O.lO . OD 254 0.05 ' 0.0: -V0 Vm L...I,JT. so 70 90 no I30 Fractions(2.5 ml / tube ) Figure 26. 90 Table l2 Recovery of Fragments from Peptide NPT Fraction lO-QM Asp % Total DNa 0 0 DNb 325 7.2 DNc I670 37 DNd 335 7.14 Percent yields are based on the amount of aspartyl residues in each fraction relative to the total aspartyl residues found in NPT prior to hydrolysis. 9] Table l3 Amino Acid Analyses of the Peptides Observed after Dilute Acid Hydrolysis of NPTa Amino Peptides ACid NPTc DNPTd DNa DNb DNc Asp 3.0 0.8 o 0.9 3.9b Thr 2.0 l.6 0 l.2 O.A Ser 3.0 3.2 0 2.8* 0.8 Glu h.6 4.h 0 2.0 0.6 Pro l.h 0 0.9 l.l Gly 2.2 2.2 0 2.0 1.2 Ala 7.0 7.0 0 3.0 2.l Val 6.0 5.6 0 2.0 0.5 lle 3.h 3.6 0 l.6* l.0 Leu 4.6 5.2 0 l.2 l.0 Lys l.8 0.6 0 2.0 0.5 Arg 2.0 2.0 0 0.8 Total hl.0 36.2 0 l02.9 l3.9 Values are expressed as residues per peptide. This value represents primarily free Asp. NPT residues/peptide before dilute acid hydrolysis. DNPT residues/peptide after dilute acid hydrolysis. This material was recovered from the t0p of the column. Amino terminus. 92 hydrophobic residues at the amino terminal portion of the peptide, only the carboxyl end of the peptide containing the seryl and aspartyl residues should be degraded. 3.ll Improved Peptide Modification Method for Sequencing. In order to keep the very hydrophobic GVP peptides in the sequencer reaction cup, the EDC-ANS method (2.l3) was used to modify the car- boxyl terminus. Figure 29-f gives the yields in logarithmic plot obtained from sequencing the synthetic tripeptide DL-Leu-Gly-Gly. Sequencing of the unmodified peptide results in total loss of the synthetic peptide to washout (Figure 29-a). The yield obtained after using the method of Foster (60) is shown in Figure 29-e. Successive treatments with EDC produces the best retention of the peptide in the reaction cup (Figure 29-9). This procedure is especially effective for insoluble peptides like NPT if 2 or more doses of EDC are given over a long time (18-h8 hours). Figure 27. 93 Peptide Modification for Improved Sequencing. The different treatments are: unmodified peptide, a; l.5 pM ANS + 0.6 pM EDC, pH 4.0, b; 0.6 pM ANS + 0.6 pM EDC + 0.l M NaCl, pH 4.0, c; 0.6 uM ANS + 0.6 pM EDC, pH 3.0, d; 0.6 pM ANS + 0.6 pM EDC, pH h.0, e; 0.3 uM ANS + l.5 HM EDC, pH 4.0, f; 0.3 pM ANS + 1.5 pM EDC) x 2, pH h.0, g; 0.3 pM peptide DL-Leu-Gly-Gly- used in all samples. Total reaction time for all samples = h hours. 9h mole-(654 3 2 22» .x. to 8. Residue number Figure 27. 95 DISCUSSION h.l General Problems Encountered in Purification, Separation, and Sequencing of GVP Peptides. Numerous difficulties were encountered in working with the GVP. Foremost among these was the problem of solubility, a feature common to both integral membrane proteins and proteins like collagen and elastin. Although attempts were made to modify the protein by maleylation, succinylation, etc., these failed. The lysine residues react with maleic anhydride only after boiling the GVP, (3.7). Lysine residues, besides being buried in the interior of the protein, as indicated by these re- sults, may be cross-linked to other portions of the protein. How- ever, trypsinization of lysine residues occurs and this makes cross-linkage rather unlikely. In addition, Jones g£_gl. (79) demonstrated that in intact gas vacuoles, lysines are accessible to spin labels. Thus, more extensive maleylation of unboiled GVP may be necessary in order to achieve modification of the lysine residues. Problems were also encountered in the area of the production of peptides. As seen above (3.7, 3.8), several methods of gener- ating peptide fragments did not work with the GVP and those which did, as tryptic digestion and NBS cleavage, produced only a fair yield of peptides. Without doubt, these problems are directly related to the insolubility of GVP. Since others (#7, 80) have failed to generate peptides from GVP even these small yields are encouraging. 96 Once peptides were made, additional difficulties presented themselves in the isolation and purification procedures. The losses during purification of peptides (Tables 7 and l0) are large. Efforts to reduce these losses have failed and it became apparent that peptides were binding irreversibly to not only ion change matrices (3.5% but also to Sephadex gels which is rather unusual (Table 7). This appears to be due to the fact that as more material is run over the same column bed, fewer sites are available for irreversibly binding to the peptides (3.5). Despite the use of many modification methods, segments of the protein remain yet to be sequenced. However, up to now this is not unusual as only one other hydrophobic protein has ever been sequenced, glycophorin (30). Moreover, the properties of glycophorin do not compare to those of GVP. Even the hydrophobic domain of glyc0phorin does not present the extreme problems of residue overlap as does the amino terminal region of the peptide NPT (3.l0). As indicated in Table 8, only two NBS peptides were obtained in yields sufficient to warrant sequence analysis. It is possible that not all of the tyrosine and tryptophan residues in GVP reacted with NBS; incomplete cleavage fragments may yet be found. Cases of unreactive tyrosine and tryptophan have been reported (8], 82, 83). h.2 An Improved Method for Reduction ofggxtractive Losses of Peptides. Since complete sequences of long peptides are seldom obtained in single runs using the sequencer, modification of the carboxyl terminal residues has become the standard method of 97 ending premature losses of residual peptide with the wash fluids. Current methods were discussed earlier (1.7), however, I have improved the carbodiimide method of Foster g£_§l. (60) (Figure 27-e) by using excessive amounts of EDC in small but frequent doses over extended periods of time. This has improved yields by 5-20% for . 2?: . E‘kiz' soluble tryptic and NBS peptides, especially for those with solu- bility problems as the peptide NPT. This extensive dose procedure seems necessary since EDC is slowly hydrolyzed by water to the .311"._ _-._ .1. I. corresponding urea. The coupling reaction should be driven to completion upon addition of excess ANS (8h). However, upon testing, this failed, as seen in Figure 27-b. The low sequencer efficiency caused by increased ANS is probably due to undesirable side reac- tions with the sequencer reagents. Whereas, increasing only the EDC concentration by means of this improved method, enhances the yields. h.3, Molecular Weight Determination of £13. Evidence is discussed here which will clarify some of the discrepancy in the reported molecular weight of GVP (hl, #3). Based on the amino acid composition of gas vacuoles from either [1. aeruginosa or A. .f_l_g§_- gauge and assuming the protein contains one tryptophan per mole of protein, a molecular weight of about 7,300 is reached (mean residue molecular weight for GVP = ll0). However, since a minimum of 89 nonoverlapping residues have been separated and partially sequenced, a molecular weight of at least 9,800 should be considered. Based 98 on this, there are probably 2 tryptOphans instead of l which would make the molecular weight about lh,600. In addition, a sequence analysis of the protein showed that greater than 80% of the amino terminus, alanine, was retrieved, based on a molecular weight of lh,300. if the molecular weight of the protein were 7,300, a yield closer to 200% (I60%) should have been obtained. If the molecular weight were 2l,500, a maximum yield of 50% should have been ob- tained. These results indicate a molecular weight of about lh,300 provided that the amino terminus of GVP was fully accessible to the i ' sequencer reagents and relatively complete cleavage was achieved. Considering all of this data, a minimum molecular weight of lh,600 is preposed for the GVP. h.h Sequence Analysis of Intact GVP and Alignment of Peptides. A customary procedure in beginning a sequence analysis of a protein is to attempt automated Edman degradation of the intact protein. Although it is unusual to sequence an entire protein in one attempt on an automated sequencer, large portions of the amino terminal sequence of many proteins have been reported. Such information greatly enhances the speed of sequence analysis. In addition to the primary purpose of accelerating the determination of the amino acid sequence, the sequencer can be used to test the purity of a particular protein preparation. Although this approach was not successful but for the first three residues of the GVP, these few residues eliminated the possibility that the GVP was not a single protein, unless a contaminating species had a blocked (e.g. acetyl- ated) amino terminal. Despite obtaining this short sequence of 99 the intact GVP, the protein remained in the reaction cup i.e. it did not wash out (3.3). Although the amino terminal sequence of the GVP agreed with that of the T2A3a peptide: Ala-Val-Glu-Lys, the possibility of formation of pyrrolidone carboxylic acid from glutamine was consid- ered as a possible cause for the sudden drop in the percentage yield. However, attempts to shorten the anhydrous acid cleavage step which causes this cyclization of glutamine did not alter the results. Besides, if the identification of T2A3ais accepted as the amino terminal peptide of GVP, then formation of pyroglutamate could not possibly be the problem. Since there are no unusual amino acids found in the GVP, it can only be suggested that there may be a tertiary protein inter- action (e.g. via an interpeptide bond) which is not readily denatured either by treatment with 88% formic acid or heat. This seems pos- sible since the protein is rather insoluble in most organic and aqueous solvents (39). Work currently in progress with Dr. J. Foster of Boston University School of Medicine should establish if any cross-links between polypeptide chains of GVP are found as in some structural proteins of eucaryotes, e.g. elastin (85). The sequence analysis of the intact GVP enabled positive identification of the peptide T2A3a as the amino terminus of the protein. In addition, it definitely showed that the NPT peptide, although beginning with the same sequence (Ala-Val) was not just a lengthy amino terminal fragment since removal of the amino l00 terminal alanine and subsequent production of peptide NPT did not alter the amino acid sequence of NPT. With the additional information from the NBS peptides that NAI was the carboxyl terminus of the protein, placement of NPT internal to the protein as seen in Figure 28 became obvious. Based on the fortuitous occurrence of the TlPlb tryptic peptide, a result of incomplete tryptic cleavage, it was also possible to align the carboxyl end of the protein. Thus, as mentioned before (3.6), TlP2 is the carboxyl terminus of TlPlb. Alignment of the other tryptic fragments T2A3b, T2A3, TZAh, and TlPla remains un- settled. Consequently, only a small portion of the peptides of the GVP can, as yet, be aligned. Those tryptic peptides shown in parenthesis in Figure 30 may be incorporated, at least in part, in the NPT peptide. Complete anal- ysis of peptide NPT may allow such conclusions. h.5 Discussion of the Peptide NPT. Twp aspects of the sequence analysis of the GVP are intriguing. These are the recovery and partial sequence of the very hydrophobic peptide NPT, and the occurrence of an octapeptide which is thrice repeated in the protein. The NPT peptide, which according to its amino acid composition contains greater than 50% hydrophobic residues, has a unique amino terminal sequence: a stretch of l5 aliphatic amino acids. As shown in Table l0, some of this sequence is questionable, although the same sequence was obtained using different peptide modification methods and both GLC and hydriodic acid hydrolysis for residue l0l identification. In any sequence analysis dealing with consecutive aliphatic residues, as residues number 2 through A (Val-Val-Val) in NPT, as much as 50% overlap can and does exist (56, 70). More- over, it is known that peptide bonds formed between two B-branched amino acids (e.g. Val-lle, Ile-Ile) are extremely difficult to hydrolyse and probably form the most acid resistant type of pep- 1 tide bond (86). For example, more than 50% of residue number 5, leucine, routinely appears with residue number 6 of peptide NPT. b However, since valine appeared at position number 6, leucine was I observed as an overlap and not as the seventh residue. Similar but as yet unresolved problems occur for residues number 3, h, 7, 9, and II. All attempts at resolving the identity of these residues using hydriodic acid have met with little or no success as the background appearance of NPT peptide washout is high enough to ob- scure the appearance of these low-yield, PTH residues. Thus, the identification of all of the NPT residues number I through l5 has relied solely on GLC analysis. If the sequence of the NPT amino terminal is as shown (Table l0), it is one of the most hydrophobic stretches of a protein yet described. The only other comparable sequences seem to be the intramembranous region of glycophorin (30) and the glycine-and alanine-rich peptides from elastin (85). These, however, do not contain such an aliphatic stretch as they do hydrophobic and neutral amino acids. Four other proteins are known which have either very hydrophobic regions or amino acid composition and which might have such aliphatic domains. l02 c_;b_3 mmn_uama (.2 we.» .uocm__m uo> no: men muoxumcn esp .m>o ecu c_ mou_uao¢ mmz tam u_ua>ch osu wo ucoscmm_< .wm oc:m_m <_z Nm_h ‘b — ..- m_< - m_< - A_m>v - Aotav - m_< - Atomv - >_a — e.g.» L A use - sms - >_a - _m> - m_< - 3.9 - m_< - L>p - mt< -Amosv_mot m_v - 0.; n_m_h - _m> - m_< - m_< - tom - x_o - tsp - so; - >_u - _m> - m_< - s_o - m_< - t>p i hmz m_m_h q<~h a_a_h _ nm<~c ~<~h - ms; - s_u - _m> - m_< b d ‘Ir— mm<~k l03 They are the hydrophobic peptide from cytochrome b5 (l9), rhodopsin (87), the liver membrane protein (88), and C55 iSOprenoid alcohol phosphokinase (89). Moreover, the last three proteins possess the highest known content of hydrophobic residues of any membrane-bound proteins (Table lh). However, no structural data are yet available 3% for any of these polypeptides. 8 Since I have established that the NPT peptide resides in the f intrapeptide region of the GVP, it suggests that the gas vacuole '1 membrane protein is following the pattern observed in glyc0phorin ' and cytochrome b5 which have their hydrophobic regions embedded in the lipid bilayer of the membrane. One may speculate that the gas vacuole membrane, having no lipid, compensates for this problem by creating its own hydrophobic environment which is interior to the vacuole surface, a feature suggested by Walsby (hl). It is known that the hydrophobic portions of some proteins, notably cytochrome c, are maintained over long evolutionary periods, although the total amino acid composition has changed up to 50% throughout this time (90, 9i). Perhaps, the gas vacuole membrane exists as a living fossil; a prototype for the more complex membranes. Since the high I degree of hydrOphobicity is apparently a common feature of all inte- I gral membrane proteins, this would provide a conservative mechanism for maintaining the integrity of membranes over long periods of time. Segrest e£_gl. (20, 30) have postulated that the hydrophobic domain of glycophorin, which has been shown to span the lipid l0h Table lh Relative Content of Polar, Intermediate, and Apolar Amino Acids for Some Membrane-Bound Proteins Membrane Protein Apolar(%) Polar(%) lntermediate(%) Ref. Chromatium Sulfur hl.7 34.2 29.0 9h Membrane (single protein) Rhodapseudomonas 50.8 27.5 20.7 95 spheroide -envelope CSS-iSOprenoid 63.0 l8.8 l8.2 39 alcohol phosphokinase Glyc0phorin (intra- 60.0 lh.5 2l.6 30 membranous region) Cyt. bS 43.3 32.5 24.2 l9 Purple membrane 55.7 23.l 22.0 48 ‘fl. halobium Gas Vacuole Membrane ‘M. aeruginosa 58.2 23.7 18.3 39 ‘A, flos-aguae 5#.l 26.9 l9.l 47 .H. halobium 52.0 28.5 l9.h 80 Proteolipid protein, 60.9 l5.6 23.2 96 peripheral myelin Proteolipid protein, 58.8 l6.8 25.3 97 heart Proteolipid protein, 56.8 l7.l 26.] 98 central w. matter Proteolipid protein, 55.4 l7.l 26.0 99 central myelin Protein N-2, central 56.] l8.l 28.2 l00 myelin Coat Protein Bacteriophage FD 63 l2 2# 50 Q Beta (+0 9 32 27 so TMV 1+2 28 29 50 105 bilayer of the erythrocyte membrane is helical. Other authors studying the CD spectra of isolated membranes, have also suggested that integral membrane proteins have a high helical content (l5). X-ray (hl), CD (AA), and IR (#6) analyses of the intact gas vacu- ole membrane protein indicate the presence of both helix and sheet structure. However, the location of these secondary structures is unknown. Based on the fact that the NPT peptide is very insoluble in aqueous solvents a sheet conformation was suspected. Moreover, the Chou and Fasman method of prediction of protein conformation (32), indicated that the NPT peptide sequence favors sheet forma- tion. Where the relative strength of helix formers is Ha>ha>la>ia> ba)Ba and the relative strength of sheet formers is Hb>hb>lb>ib> bb>Bb; (Pa) and (Pb) are the average amino acid residue strengths of formation of helix and sheet, respectively (32): helix Ha ha ha ha Ha ha ha Ia sequence Ala-Val-(Val)-(Val)-Leu-Val-(Val)-Ile- sheet lb Hb Hb Hb hb Hb Hb Hb Ia Ha Ha Ha Ha ha Ia 1.20 =(Pa) continued (Ile)-Leu-(Leu)-Ala-(Leu)-(Val)-(Ile) Hb hb hb lb hb Hb Hb 1.1.3 =(Pb> Although the hydrophobic domain of the GVP does not have a helical nature, this does not preclude the possibility of helix existing elsewhere in the protein. 4.6 A Repeating Octapeptide in the GVP. The second intriguing aspect of the primary structure of the GVP is the presence of a 106 repeating octapeptide as found in peptides TlP2 and NIA, TlPla, and TlPlb (Tables 7 and ID): Ala-GIu-Ala-Val-Gly-Leu-Thr-Glu Since these peptides have been shown to have different elution patterns, different amino termini (serine or tyrosine), and differ- ent carboxyl termini (alanine, arginine, lysine), they are defin- itely separate peptides; that is, they are not merely different overlapping fragments of one sequence region. ' Repetitive sequences are common to structural proteins such as collagen and elastin (50). Many investigators, among them Nolan and Margoliash (#9), have suggested that the presence of repeating amino acid sequences in the primary structure of some proteins like immunoglobins might indicate that ”partial gene duplications repre- sent a common and important evolutionary mechanism for increasing the size of polypeptide chains”. This may well be the case for immunoglobins, ferredoxins, and clupeines as they suggest. However, in structural proteins, sequence repetition appears to be more important as a structural building block rather than as a method for increasing chain length. For example, in elastin the short- range repetition of sequences (Pro-Gly-Val-Gly-Val-Ala-)n and (Pro-Gly-Val-Gly-Val)n suggest a regular structure. Thus, the “oiled-coil“ model was proposed for elastin in which these repeating units are theorized to account for the extensibility of the protein (85). Perhaps, this periodicity of sequence in the GVP accounts for its unique rigidity. Based on the Chou and Fasman method (32) of l07 prediction of secondary structure, these repetitious regions should have a strong tendency to form helix as shown below: helix Ha Ha Ha ha Ba Ha ia Ha = l.22 (Fa) sequence Ala-Glu-Ala-Val-Gly-Leu-Thr-Glu sheet Ib Bb lb Hb ib hb hb Bb=0.92 (Pb) Underlining denotes the predicted helical regions for each of these peptides: TlPlb Tyr-Ala-Glu-Ala-Val-Gly-Leu-Thr-Glx-Ser-Ala-Ala-Val-Pro-... TlPla Ser-Ala-Glu-Ala-Val-Gly5Egu-Thr-Glx-Val-Ile-Ala-... TlP2 Tyr-Ala-Glu-Ala-Val-Gly1Leu-Thr-Glx-Ser-Ala-Pro-Val-Ala-Ala. If indeed these are periodic, helical regions in the protein, then what could lend greater strength to a structure than a series of coils or l'cylinders", the geometry of which represents one of the strongest structures available to architectural design. Only the complete sequence of the GVP can show if more of these repetitive segments exist in the protein. As stated above, predictions, based solely on primary structure, can be made regarding the helical, sheet, turn and coil regions of a particular polypeptide with greater than 85% certainty (32). Appli- cation of this method of prediction of secondary structure based only on peptide fragments tends to underpredict the amount of helix and sheet in a protein since it neglects to consider the tertiary folding of the protein which can provide stability to secondary structures (32). Nevertheless, the method can provide valuable information for a proposed structural model of the molecule. All of the other GVP A“ 108 peptides which have been sequenced are too short to provide enough information for a prediction of secondary conformation with the exception of T2A2: helix Ba Ia ha Ia ia Ha Ha ia=l.0’+ sequence Gly-Ile-Val-lle-(Asp)-(Ala)-Ala-Arg sheet lb Hb Hb Hb lb lb lb lb=l.lS(Pb) This peptide can be evaluated as probably having either a coiled or helical configuration but not sheet structure as it contains charged ' residues in a sheet nucleation area (32) which makes sheet conforma- tion less favorable. h.7 Implications of the Amino Acid Composition of GVP end Its Relationship to Other Integral Membrane Proteins. Much information can be gleaned from a comparison of the amino acid composition of proteins which are considered to be related in function or structure. Not only can relative polarities (22, 92) be determined from the amino acid composition, but also calculations of relative amounts of secondary structure (93). The results of the improved amino acid analysis (Table h) show that valine, leucine, and isoleucine constitute 35% of the residues of the GVP of Microcystis aeruginosg, a value close to those found for GVP from Anabaena flos-aguae (32%) (#7) and Halobacterium halobium (33%) (80). It is of interest that the gas vacuole membrane protein from bacteria and blue-green algae all contain over 50% l09 hydrophobic residues as based on the criteria of Tanford (92) and Capaldi and Vanderkooi (22).3 The relative percentages of hydrophobic, polar and intermediate groups in various integral membrane proteins fall into a distinct category (Table lh). If compared to corresponding values for non- membrane-bound or peripheral proteins, the polarity, as defined in Footnote 3, of most integral membrane proteins is between 29 and h0% for polar amino acids. The corresponding values for most periph- eral proteins fall between hi and 53% (22). From such data the relationship emerges that the less a particular protein is associ- ated with a membrane structure, the more polar it becomes with a concomitant decrease in hydrophobicity. Such a relationship has been alluded to by others (I7, 22) but, perhaps, the most fascin- ating insight is the similarity between those integral proteins from procaryotes and those from eucaryotes; they appear the same in relative polarity. Similarities as those described above emphasize that the gas vacuole membrane protein is indeed a true membrane protein of the integral-type. Although the gas vacuole membrane may be said to be only a half-membrane in the morphological sense that it contains no lipid, it does have an integral-type membrane protein. 3 This operational definition divides amino acids into three classes (92, 22): Polar (Asp, Asn, Glu, Gln, Lys, Arg); Intermediate (Ser, Thr, His, Gly); Nonpolar (Ala, Val, Ile, Leu, Cys, Met, Pro, Phe, Trp, Tyr).' llO Krigbaum et_gl. (93) have devel0ped a procedure whereby the amount of secondary structure in a globular protein may be predicted from its amino acid composition. The average errors of this method are 8.2% for helix, 8.2% for sheet, 5.5% for turns, and 5.7% for coil regions. Regions of coil and turn are as defined in Krigbaum _g£_§l. (93). The error is even less if the acid and amide side chains can be distinguished. I applied this empirical approach to the GVP. The following sets of relations were used in predicting the percent of secondary structure in the GVP based on the amino acid composition as given in Table A; where Ha+ and HB+ are helix formers; HA- and H3- are nonhelix formers; the same holds for sheet (3*, a"), turn (r+, T'), coil (c+, c-) (93): helix: HA+ = Ala + Leu + His + Ile HA' = Pro + Thr + Try + Met + Tyr HB+ = Ala + Leu + His + Tyr HB' = Pro + Thr + Ile + Val + Arg sheet: B+ = Asp + Thr + Arg + Pro + Val + Asn B' = Leu + Ala + Glu + Gln + Gly turn: T+ = Gly + Thr + ASp + Gln + Glu + Met + Asn T = Ser + Ala + Phe + His + Cys coil: C+ = Cys + Tyr + Ala + Thr C = His + Asp + Asn % helix = h3.02 + 0.7.07 HA+ + 0.676 HB+ - 1.223 HA“ - 0.865 H8 % sheet = 19.13 + 1.633 3* - l.u77 e' lll % turn = 25.91 + 0.9ou 7+ - 0.909 T' % coil = 2h.l9 + 1.125 c+ - 2.041 c- Using this set of equations predicts the GVP to contain 35.5% helix, 20.2% sheet, 25.5% turn, and l5.5% coil. Whether or not these calcu- lations are valid for a hydrophobic membrane protein depends on whether comparisons can be made between integral membrane proteins and glob- ular proteins. X-ray data for many more proteins are necessary before such assertions can be completely validated. In addition, estimation of secondary structure in the GVP has not yet been fully evaluated by other physical methods such as X-ray, CD and IR analyses (See Introduction (l.5) for current available data.). Such information, ought to indicate whether or not this prediction method is valid for nonglobular proteins. ‘ h.8 Tentative Molecule; Model for the Function of the Gas Vacuole Membranes. Based on the above data, it is possible to propose a functional model for the gas vacuole membrane. Gas vacuoles appar- ently do not actively regulate the flow of gas through the vacuole membrane; that is, gases appear to pass through the membrane by diffusion (#0, lOl). Since many diatomic gases are apolar, passage through an apolar milieu would provide a path of least resistance. This would also be the case for other gases as Ar, CH“, and C02 which have been shown by Walsby to freely permeate the membrane (#1). Such an apolar milieu would exist in the aliphatic, amino terminal portion of the peptide NPT. Movement of gases through the membrane llZ is dependent on several factors. Initially, it is necessary to assume that a subunit substructure exists in the membrane as already proposed (hl, l02). Secondly, in order to allow gas to move through the membrane without a conformational change in the GVP (for which there is no evidence), either the gas must pass through the inter- molecular space of the protein, or through pores. The lining of either of these passageways would be the aliphatic portion of the NPT peptide. Polar molecules would, therefore, be excluded from passing through this region. Larger apolar molecules would also be restricted based on the rigidity and size of the space which must be at least 3 A in order to accomodate molecules as large as Nzand CH4 for which the van der Waals radii are known to be about 3.0 o and 2.5 A, respectively. BIBLIOGRAPHY .P'UJ VO‘U‘I 12. 13. 1h. '5. 16. I7. BIBLIOGRAPHY Lenard, J. and Singer, S. J. Proc. Nat. Acad. Sci. 56, 1828 (1966). Davson, H. and Danielli, J. F. J. Cell Physiol. 5, A95 (1935). Robertson, J. D. Progr. Biophys. Chem. 19, 343 (I960). Korn, E. D. Science, 123 lh9l (I966). Kavenau, J. L. Fed. Proc. 25, 1096 (1966). Stockenius, W. and Engelman, D. M. J. Cell Biol. 32. 613 (1969). Engelman, D. M. J. Molec. Biol. 28, 153 (1971). Singer, 5. J. Ann. Rev. Biochem. ‘33, 866 (197h). James, R. and Branton, D. Biochem. Biophys. Acta 323, 378 (1973). Scott, R. E. and Carter, R. L. Nature New Biol. 233. 219 (I971). "'- Dupont, Y., Harrison, S. C. and Hasselbach, W. Nature 253, 555 (I973). Wallach, D. F. and Gordon, A. Fed. Proc. 21, 1263 (I971). Chuang, T. F., Awasthi, Y. C., Funk, L., Crane, F.L. Biochem, Biophys. Acta 211, 599 (1970). Wallach, D. F. J. Gen. Physiol. 23, 35 (I969). Lenard, J., and Singer, S. J., Proc. Nat. Acad. Sci. 26, 1828 (1966). Wallach, D. F. and Zahler, P. H. Proc. Nat. Acad. Sci. 26, 1552 (1966). Singer, S. J. and Nicolson, C. L. Science 175, 720 (I972). 114 18. 20. 21. 22. 23. 2#. 25. 26. 27. 28. 29. 30. 31. 32. 33. 3#. 115 Spatz, L. and Strittmatter, P. J. Biol. Chem. 238, 793 (I973). Spatz, L. and Strittmatter, P. Proc. Nat. Acad. Sci. '68, IO#2 (1971). Segrest, J. P., Kahane, 1., Jackson, R. L. and Marchesi, V. T. Arch. Biochem. Biophys. 155, 167 (1973). Salton, M. R.: in ”Biomembranes,“ (ed. Manson, A.). New York: Plenum Press. (1971) p. #0. Capaldi, R. A. and Vanderkooi, G. Proc. Nat. Acad. Sci. .69, 930 (I972)- Crane, L. J. and Lampen, J. 0. Arch. Biochem. Biophys. 169, 655 (l97#) . Tanford, C.: in ”The Hydrophobic Effect: Formation of Micelles and Biological Membranes," New York: John Wiley and Sons. (1973). Scheraga, H., Nemethy, G. and Steinberg, I. J. Biol. Chem. 231, 2506 (I962). Bigelow, C. J. Theoret. Biol. 16, 187 (1967). Singleton, R. and Amelunxen, R. E. Bact. Rev. 31, 320 (I973) Goldsack, D. E. Biopolymers 9, 2#7 (1970). Bull, H. B. and Breese, K. Arch. Biochem. Biophys. 158, 681 (I973). "" Segrest, J. P., Jackson, R. L. and Marchesi, V. T. Biochem. Biophys. Res. Commun. #2, 96# (1972). Marchesi, V. T., Tillack, T. W., Jackson, R. L., Segrest, J. P. and Scott, R. E. Proc. Nat. Acad. Sci. '69, 1##5 (1972). Chou, P. Y. and Fasman, G. 0. Biochem. .13, 222 (197#). Kotelchuck, D. and Scheraga, H. A. Proc. Nat. Acad. Sci. 62, l# (1969). Lewis, P. N., 66, N., 06, M., Kotelchuck, D. and Scheraga, H. A. Proc. Nat. Acad. Sci. .65, 810 (1970). 35. 36. 37. 38. 39. #0. #I. #2. #3. #5. A6. A7. #8. #9. 50. 51. 52. 53. 116 Finkelstein, A. V. and Ptitsyn, O. B. J. Molec. Biol. 6;, 613 (1971). Chou, P. Y. and Fasman, G. D. J. Molec. Biol. 13, 263 (I973). Robson, B. and Pain, R. H. J. Molec. Biol. 66, 237 (I971). Puett, 0. Biochem. Biophys. Acta £61, 537 (1972). Jones, D. D. and Jost, M. Arch. Mikrobiol. .16, #3 (1970). Walsby, A. E. (Proc. Roy. Soc. - Ser. 3 118, 301 (1971). Walsby, A. E. Bact. Rev. _36, l (1972). Waaland, J. R., Waaland, S. D. and Branton, D. J. Cell Biol. 66, 212 (1971). Weathers, P. J. unpublished data. Weathers, P. J., Lalitha, S. and Haug, A. unpublished data using CD showed 2#% helix, 36% sheet, and #O% coil present in inflated gas vacuole membranes. Jones, D. D. Ph.D. thesis, Michigan State University (1970). Jones, D. D. and Jost, M. Planta 100, 277 (I971). Falkenberg, P., Buckland, B. and Walsby, A. E. Arch. Mikrobiol. g5. 30!» (I972). Stockenius, W. and Kunau, W. H. J. Cell Biol. ‘36, 337 (1968). Nolan, C. and Margoliash, E. Ann. Rev. Biochem. {21, 727 (1968). Dayhoff, M. 0.: in ”Atlas of Protein Sequence and Structure,” Washington, D. C.: National Bio Medical Research Foundation. (l968-l972). Wu, T. T., Fitch, W.M. and Margoliash, E. Ann. Rev. Biochem. 5;. 539 (1971+). Liljas, A. and Rossman, M. G. Ann. Rev. Biochem. 56, #75 (1971+). Tschesche, H. and Kupfer, 5. Eur. J. Biochem. 26, 33 (I972). 5#. 55. 56. 57. 58. 59. 60. 61. 62. 63. 6#. 65. 66. 67. 68. 69. 117 Callahan, P. X., McDonald, J. K. and Ellis, S. Fed. Proc. 3_1. 1105 (1972). Paukovits, W. R. J. Chrom. 63, 15# (I973). Edman, P.: in “Molecular Biology, Biochemistry and BiOphysics: Protein Sequence Determination,“ (ed. Needleman, S. D.) New York: Springer-Verlag (1970). Edman, P. and Begg, G. Eur. J. Biochem. l, 80 (I967). Inman, J. K., Hannon, J. E. and Appella, E. Biochem. Biophysics. Res. Commun. #6, 2075 (1972). Braunitzer, G. Schrank, B. and Ruhfus, A. Hoppe-Seyler's Z. Physiol. Chem. 351, 1589 (1970). Foster, J. A., Bruenger, E., Hu, C. L., Albertson, K. and Franzblau, C. Biochem. Biophys. Res. Commun. 33, 70 (I973). Wachter, E., Machlerdt, W., Hofner, H. and Otto, J. FEBS Lett. 35. 97 (I973). Previero, A., Derancourt, J. and Coletti-Previero, M-A. manuscript in preparation. Gorham, P. R., McLachlin, J., Hammer, U. T. and Kim, W. K. Verh. int. Ver. theor. angew. Limnol. 13, 796 (196#). Lowry, O. H., Rosebrough, N. J., Farr, A. L. and Randall, R. J. J. Biol. Chem. 193, 265 (1951). Gray, W. R.: in “Methods in Enzymology,“ (ed. Hirs, C. H. W.) Vol. 11 p. 139. New York: Academic Press (1967). Schroeder, W. A., Jones, R. T., Cormick, J. and McCalla, K. Anal. Biochem. 35, 1571 (1962). Butler, P. J. G., Harris, J. I., Hartley, B. S. and Leberman, R. Biochem. J. 112, 679 (1969). Ramachandran, L. K. and Witkop, B.:- in ”Methods in Enzymology,“ (ed. Hirs, C. H. W.) Vol. II p. 290. New York: Academic Press (1967). Offord, R. E. Nature 11 591 (1966). 70. 71. 72. 73. 7#. 75. 76. 77. 78. 79. 80. 81. 82. 830 8#. 85. 86. 87. 118 Foster, J. A. personal communication. Schultz, J.: in “Methods in Enzymology,“ (ed. Hirs, C. H. W.) Vol. II p. 255 New York: Academic Press (1967). Fletcher, J. C. Biochem. J. 68, 3# c (1966). Moore, 5., Spackman, D. H. and Stein, H. H. Anal. Chem. 39, 1185 (1958). Schmidt, P. 1.: in “Techniques in Amino Acid Analysis,” Chertsey: Technicon Instruments Company Ltd. (1966). Beckman Model 890c Sequencer Instruction Manual, January 1972. Pisano, J. J. and Bronzert, T. J. Anal. Biochem. ‘33, #3 (1972). Smithies, 0., Gibson, 0., Franning, E. M., Goodfliesh, R. M., Gilman, J. D. and Ballantyne, D. L. Biochem. .1Q, #912 (I971). Inglis, A. 5., Nicholls, P. W. and Roxburgh, C. M. Aust. J. Biol. 35, 12#7 (1971). Jones, D. D., Haug, A., Jost, M. and Graber, D. R. Arch. of Biochem, 8i0phys. I35, 296 (1969). Krantz, M. J. and Ballou, C. E. J. of Bact. 115, 1058 (I973). Khandwala, A. S. and Kasper, C. 8. Biochem. Biophys. Acta 231. 348 (1971)- Gross, E. and Witkop, B. J. Biol. Chem. 231, 1856 (I962). Smyth, D. G., Stein, W. H. and Moore, S. J. Biol. Chem. 231, l8#5 (1962). Carruway, K. L. and Koshland, D. E.: in “Methods in Enzymology,“ (ed. Hirs, C. H. W. and Timasheff, S. N.) Vol. 25 p. 616 New York: Academic Press. (1972). Gray, W. R., Sandberg, L. B. and Foster, J. A. Nature 2#6, 55' (I973). Kasper, C. 8.: in “Molecular Biology Biochemistry and Biophysics: Protein Sequence Determination,” New York: Springer-Verlag. p. l#2 (1970). Vanderkooi, G. and Sundarlingam, M. Proc. Nat. Acad. Sci. .61, 233 (I970. 88. 89. 90. 91. 92. 93. 9#. 95. 96. 97. 98. 99. 100. 101. 102. 119 Hinman, N. D. and Phillips, A. H. Fed. Proc. ‘31, Abs. 1086. Sanderman, H., Jr. and Strominger, J. L. Proc. Nat. Acad. Sci. §§, 2541 (1971). Perutz, M., Kendrew, J. and Watson, H. J. Mol. Biol. 13, 669 (1965). Margoliash, E. and Schejter, A. Advan. Prot. Chem. .11, 113 (1966). Tanford, C. J. Am. Chem. Soc. .95: #2#O (I962). Krigbaum, W. R. and Knutton, S. R. Proc. Nat. Acad. Sci. 19, 2809 (1973). Schmidt, 0. L., Nicolson, G. L. and Kamen, M. 0. J. Bact. .lQfi. 1137 (I971)- Niederman, R. A., Segen, B. J. and Gibson, K. D. Arch. Biochem. Biophys. 152, 5#7 (1972). Eng, L. F., Chao, F. C., Gerstl, 8., Pratt, D. and Tavastsjerna M. 6. Biochem. .1, #455 (1968). Eichberg, J. Biochem. BiOphys. Acta 181, 533 (1969). Tenembaum, D. and Folch-Pi, J. Biochem. BiOphys. Acta 115, 141 (1966). Wolfgram, F. and Rose, A. S. J. Neurochem. .8, 161 (1961). Gagnan, J., Finch, P. R., Wood, D. D. and Moscarello, M. A. Biochem. 1_8_, #756 (1971). Walsby, A. E. Proc. Roy. Soc. Ser. B 113, 235 (1969). Jost, M. and Jones, D. D. Can. J. Microbiol. ‘16, 159 (I970). III III 8 16 IIIIIIIIII I 93 031 31 I 1