Proc. Nati. Acad. Sci. USA Vol. 75, No. 8, pp 3708-3712, August 1978

Biochemistry

Lysyl-derived aldehydes in outer membrane proteins of Escherichia coli (allysine/a-aminoadipic acid -semialdehyde)

DANA L. DIEDRICH* AND CARL A. SCHNAITMANt Department of Microbiology, University of Virginia School of Medicine, Charlottesville, Virginia 22901

Communicated by Albert L. Lehninger, June 2,1978

MATERIALS AND METHODS Strains, Media, and Isotopes. E. coil K-12 W1485F- was grown on L-salts medium with 0.5% potassium lactate as a carbon source (5). Outer membrane proteins were labeled with [U-14C]lysine by growing cultures in the salts medium supplemented with 20 amino acids plus vitamins (5). Carrier lysine was at 25 gg/ml, and radioactive lysine (specific activity of 312 Ci/mol) was used at 0.1 uCi/ml of culture medium. Purification of Outer Membrane Proteins. The major outer membrane proteins were purified as described (5), except that the ion-exchange chromatography step was omitted. Polyac-

rylamide gel electrophoresis was performed as described (5, 7). Treatment with Phenyihydrazine. Two milligrams of purified major outer membrane protein 1 in 200 Al of pH 6.6 phosphate buffer (0.1 M) containing 0.1% (wt/vol) sodium dodecyl sulfate (NaDodSO4) was mixed with 20 Al of phenylhydrazine in 70% ethanol to a final concentration of 20 mM. This was incubated at room temperature for 6 hr and then passed over a Bio-Gel P2 column equilibrated against the same buffer. A difference spectrum of the reacted protein against the unreacted protein was obtained with a Cary 14 recording spectrophotometer from 250 to 400 nm. NaB3H4 Reduction of Proteins. Purified outer membrane proteins were reduced with NaB3H4 (in pH 7.4 phosphate buffer containing 0.1% NaDodSO4) at concentrations of 1-5 mg/ml. NaB3H4 was dissolved in 50 mM NaOH immediately prior to use, and it was added to slowly stirred samples of protein at room temperature in five increments at 15-min intervals. Protein samples received 1-5 mCi of NaB3H4 per mg of protein at specific activities ranging from 60 to 280 Ci/mol. Nonradioactive borohydride was added in two 15-min intervals in order to chemically complete reduction. Generally, each protein sample received 33Amol of NaB3H4. The pH of the reaction was monitored, and 50 mM HCl was used to maintain the pH when necessary. The reaction was terminated by adding HCl until the pH dropped to approximately 2. The protein was then precipitated with 3 vol of acetone. The precipitate was dissolved in a small volume of NaDodSO4-containing phosphate buffer and subsequently boiled for 3 min. The protein was passed over a Bio-Gel P2 column and examined on polyacrylamide gels to assess the extent of 3H incorporation and to determine if reductive cleavage of the peptides had occurred. Oxidation of Outer Membrane Proteins. Performic acid oxidation was performed according to Miller et al. (9). Lyophilized, NaDodSO4-free, purified outer membrane proteins were dissolved in performic acid at 1-5 mg/ml, and oxidation was carried out in an ice bath for 2 hr. The oxidized protein was diluted 1:10 in water and lyophilized. Reaction of Protein with Cyanide and Ammonia. Five milligrams of [U-14C]lysine-labeled purified outer membrane protein 3a, in 3 ml of H20 (containing 2% NaDodSO4), was mixed with 3 ml of 30% NH4OH containing 300 mg of NaCN. This was incubated at room temperature with slow stirring for 2 hr. The reaction was terminated by the addition of HC1 until the pH dropped to approximately 2. The protein was precipitated by the addition of 3 vol of acetone, and the protein precipitate was washed twice with acetone/water (6:1 vol/vol) and then lyophilized.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisemewnt" in accordance with 18 U. S. C. §1734 solely to indicate this fact.

Abbreviation: NaDodSO4, sodium dodecyl sulfate. * Present address: Department of Microbiology, Texas Tech University School of Medicine, P.O. Box 4569, Lubbock, TX 79409. t To whom reprint requests should be addressed.

ABSTRACT The major outer membrane proteins from Escherichia coli K-12 are modified to contain a-aminoadipic acid 5-semialdehyde (allysine). Tile allysine was found to be derived from lysine and it was identified by derivatizing it to chloronorleucine by reduction, a-aminoadipic acid by oxidation, and to a,e-diaminopimelic acid by reacting it with CN- and NH3. The a-aminoadipic acid was identified by mass spectremetry. Two major outer membrane proteins were found to possess allysine, a modified lysine characteristically found in connective tissue. The multilayered envelope of Escherichia colh is composed of a cytoplasmic membrane, a peptidoglycan or murein layer, and an outer membrane. The outer membrane is characterized by the asymmetry of its phospholipid and lipopolysaccharide components (1, 2), and by its comparatively simple protein composition. Approximately 80% of the protein mass of the outer membrane can be accounted for by two major polypeptides, designated by Schnaitman (3) as proteins 1 and 3a. Although it has been shown that protein 1 can be resolved into two bands (la and lb) when certain polyacrylamide gel electrophoresis procedures are used (4), we (5) and others (6) have shown that the primary structure of proteins la and lb is probably identical. These data led us to suspect that proteins la and lb are identical proteins that had been subjected to post-translational modification(s) (7). We set out to chemically identify the putative post-translational modification of proteins la and lb. This led to the discovery of another modification of these proteins which also appears to be present in protein 3a and in the phage-directed protein 2 (5, 8). This communication describes the presence of allysine (a-aminoadipie acid 5-semialdehyde) in the major outer membrane proteins from E. coli. The presence of this lysylderived aldehyde has been confined to the connective tissue proteins collagen and elastin, and it serves, in part, to form the chemical crosslinks in these proteins. The potential significance of allysine in bacterial membrane proteins is discussed.

3708

Proc. Natl. Acad. Sci. USA 75 (1978)

Biochemistry: Diedrich and Schnaitman In another experiment, run in one-third the volume of the previous experiment, 125 ,Ci of Na14CN (60 Ci/mol) was used without added carrier. The proteins 3a and 1 were labeled [4,5-3H]leucine, and after the reaction was terminated, the protein was passed over a Bio-Gel P2 column to remove residual 4C radioactivity. Split-Stream Amino Acid Analysis. Protein samples were hydolyzed as described (5). Protein hydrolysates were applied to a Beckman 121 automated amino acid analyzer equipped with a stream-splitting device adjusted to supply 70% of the sample to a fraction collector and 30% of the sample to the ninhydrin detector. Later samples were also run on a Beckman 119 automated amino acid analyzer with a stream-splitting device that delivered 90% of the sample to the fraction collector and 10% of the sample to an external fluorogenic detector that used o-phthalaldehyde. Thin-Layer Chromatography. Desalted material suspected of being a-aminoadipic acid was collected from the split-stream amino acid analyzer and analyzed on cellulose thin-layer chromatography sheets in butanol/acetic acid/water (12:3:5 vol/vol). Authentic a-[14C]aminoadipic acid (Amersham/ Searle) was run as the standard. Radioactivity was localized on the chromatogram sheets by cutting the chromatograms into 5-mm strips and determining their radioactivity in a liquid scintillation spectrometer. Gas-Liquid Chromatography and Mass Spectrometry. The material suspected of being a-aminoadipic acid was collected from the split-stream amino acid analyzer and desalted on a Bio-Rad AG 50-X4 (H+) column by the method of Franzblau et al. (10). The desalted material was lyophilized, extracted twice with diethyl ether, and blown dry with nitrogen. The putative a-aminoadipic acid and an authentic standard (Sigma) were derivatized to their N-trifluoroacetyl n-butyl esters by the method of Roach and Gehrke (11). The derivatives were dissolved in ethylacetate and injected into a Finnigan gas chromatograph using a column of OV-17 siloxane substrate on a support of Chromosorb W (80-100 mesh). The effluent from the gas chromatograph was examined with a mass spectrometer adjusted to provide a 20 atomic mass unit scan of the molecular weight region of the derivative. Scans were taken over a 20-sec interval, which included the retention time of the standard derivative. Pulsed positive negative ion chemical ionization mass spectra (12) were obtained with a Finnigan Model 3300 quadrupole mass spectrometer that had been modified to simultaneously record both positive and negative chemical ionization mass spectra.

RESULTS Studies on the influence of culture medium on the production of proteins la and lb (7) suggested that these proteins might be modified by addition of a small molecule derived from intermediary metabolism. The proteins were examined for the presence of an a-keto acid prosthetic group based on the discovery of pyruvate bound to adenosylmethionine decarboxylase (13) and the finding (14), later retracted (15), of a-ketobutyrate bound to urocanase. Purified outer membrane protein 1 was reacted with phenylhydrazine in an attempt to detect a carbonyl-containing prosthetic group. A difference spectrum of the reacted protein against unreacted protein revealed a phenylhydrazone absorbance at 293 nm. Based on this finding, protein 1 was reduced with NaB3H4 and examined on polyacrylamide gels for the incorporation of 3H radioactivity. This protein, which was derived from a culture labeled with [14C]leucine, was heavily labeled with 3H (data not shown). When the labeling was carried out under

3709

prolonged conditions (pH 8.6 for 3 hr) some cleavage of the protein occurred, as indicated by the presence of both uncleaved protein and several peaks of lower apparent molecular weight. However, the ratio of 3H to 14C was identical in the uncleaved protein and in the smaller fragments. When protein was labeled under milder conditions (pH 7.4 for 1.5 hr), a single, heavily 3H-labeled peak was observed which migrated with the same mobility as unreacted protein 1. These observations, together with the inability to detect characteristic 3H-labeled a-amino alcohols in hydrolysates of the labeled protein, indicated that the bulk of the 3H incorporation could not be accounted for by nonspecific reductive cleavage, as has been reported in other systems (16). A purified preparation of ['4C]leucine-labeled, NaB3H4reduced protein 1 was cleaved with CNBr and the peptides were analyzed on gels (5). At least three of the five CNBr peptides contained 3H radioactivity, indicating that the labeling was not confined to a single site on the polypeptide chains (data not shown). Purified preparations of [U-'4C]lysine-labeled protein 1 were then reduced with NaB3H4 and hydrolyzed in 6M HC1 for 18 hr. The resulting amino acids were converted to their fluorescent dansyl derivatives and examined on polyamide thinlayer chromatograms (17). The 3H radioactivity was found predominantly in a fluorescent spot that also contained 14C radioactivity (data not shown). These results suggested that the reduced moiety in the protein was a derivative of lysine, with the modification existing as a prosthetic group attached to lysine by a linkage recalcitrant to acid hydrolysis, or the modification was in the carbon skeleton of lysine itself. The obvious precedent for the latter alternative is the presence of a-aminoadipic acid 5-semialdehyde (allysine) in connective tissue. We then hydrolyzed NaB3H4-reduced, [U-14C]lysine-labeled outer membrane protein 1 and subjected the hydrolysate to analysis on an amino acid analyzer equipped with a streamsplitting device. Fig. 1 shows the profile of the radioactivity eluted from the long column of the analyzer. The profile is similar to the results obtained when hydrolysates of collagen and elastin are run on a split-stream amino acid analyzer (18, 19). One peak containing 3H and 14C radioactivity eluted from the analyzer at the calibrated position of chloronorleucine. This would be expected since a-aminoadipic acid 5-semialdehyde is reduced by NaBH4 to hydroxynorleucine, which in turn is converted extensively to chloronorleucine upon hydrolysis in

7

0

0 I-

x

x

Ea

Ea

0

._

0

4.

._

._

0 ._

a 4) C"0 S-

Fraction

FIG. 1. Purified preparation of [U-14C]lysine-labeled protein 1 was reduced with NaB3H4 and hydrolyzed for amino acid analysis. The effluent from the long column of the split-stream amino acid analyzer was collected and radioactivity was measured by liquid scintillation spectrometry. (-) 3H radioactivity from NaB3H4; (-- -) 14C radioactivity from lysine. -ClNle, chloronorleucine.

3710

Biochemistry: Diedrich and Schnaitman

Proc. Natl. Acad. Sci. USA 75 (1978)

A

B I,

I' -_ .S-

2-AAD

C

D

HCl (18). The unidentified peak near leucine also has been observed in collagen hydrolysates (18), and the unresolved 3H radioactivity seen in fractions 60-80 is characteristic of reduced aldol (18). Reduced aldol is the reduced condensation product of two allysine residues and it is unstable to acid hydrolysis. No 3H radioactivity was found on the short column of the amino acid analyzer (data not shown). The data presented in Fig. 1 were suggestive of the presence of allysine in the major outer membrane protein 1 from E. coll; however, the resolution of the material made isolation and purification of reduced material unsatisfactory for further chemical characterization. We then chose to generate another derivative allysine that we could isolate free of substantial contamination by nonradioactive amino acids. Studies on connective tissue have shown that allysine can be oxidized to a-aminoadipic acid (9). This derivative is stable, elutes from the analyzer between alanine and valine, and is obtainable commercially in radioactive form. Purified preparations of [U-14C]lysine-labeled protein 1 were oxidized with performic acid, hydrolyzed, and analyzed on the split-stream amino acid analyzer. Fig. 2A shows the elution position of the amino acids from the long column of the analyzer; Fig. 2B shows the calibrated position for a-aminoadipic acid. Fig. 2C shows that protein isolated from a midlogarithmic phase culture possessed lysine-derived radioactivity in precisely the position of a-aminoadipic acid. Fig. 2D shows the results of a similar experiment in which the protein was isolated from a late logarithmic phase culture. In addition to the apparent presence of a-aminoadipic acid, a significant amount of lysine-derived radioactivity was present in another;peak eluting with aspartic acid. This material was present if the protein was not oxidized (Fig. 2E); however, little or no a-aminoadipic acid was found. Amino acid analysis of the radioactive lysine used to label the cultures showed no radioactive component other than lysine. The material eluted from the short column of the analyzer showed no 14C radioactivity other than lysine (data not shown). Fig. 3 shows the radioactive profile of an oxidized hydrolysate of major outer membrane protein Sa. This is another major protein, which is the so-called heat-modifiable protein (20) and has been shown to play a role in recipient ability during Fa-Aminoadipic acid Asp Glu

(ASP)

E

I

Val

50

60

E 6o,

U 40

3X3-

'0

FIG. 2. Split-stream amino acid analysis of performic acid-oxidized preparations of outer membrane protein 1. (A) Calibrated elution position for the amino acids on the long column of the analyzer. (B) Calibrated elution position for a-aminoadipic acid. (C) Profile of [U-14C]lysine-derived radioactivity eluting from the analyzer. Protein was isolated from a midlogarithmic phase culture. (D) Profile of [U-14C]lysine-derived radioactivity eluting from the analyzer. Protein was isolated from a late logarithmic phase culture. (E) Profile of [U-14C]lysine-derived radioactivity eluting from the analyzer. Protein was from a late logarithmic phase culture and was not oxidized prior to hydrolysis. 2-AAD, a-aminoadipic acid.

2-J

10

20

30

40

70

80

90

Fraction

FIG. 3. Split-stream amino acid analysis of hydrolysates of [U-l4C]lysine-labeled outer membrane protein 3a that had been oxidized prior to hydrolysis. The calibrated positions for a-aminoadipic

acid and three amino acids are indicated.

Biochemistry:

Diedrich and Schnaitman

Proc. Natl. Acad. Sci. USA 75 (1978)

mediated conjugation (21). This protein was taken from a late logarithmic phase culture, and the presence of a-aminoadipic acid is indicated. However, little radioactivity is present in the aspartate region. Considerable variation has been observed in the fraction of total lysine radioactivity present in a-aminoadipic acid after oxidation. Protein 1 possesses 1-5% of the lysine radioactivity in a-aminoadipic acid, whereas protein 3a appears to have 6-8% of the total. Program adjustments were made on another amino acid analyzer that used a single column in order to obtain sufficient amounts of a-aminoadipic acid free of significant contamination by nonradioactive amino acids. Samples of the putative a-aminoadipic acid and the material eluting with aspartate from both proteins 1 and 3a were collected from the analyzer and desalted. These samples were analyzed by thin-layer chromatography on cellulose and compared to the chromatographic characteristics of authentic a-[14C]aminoadipic acid. The results (Fig. 4) show that the putative a-aminoadipic acid runs with the same RF as the authentic standard, and the material that runs with aspartate is only detected in hydrolysates from protein 1. These results were also obtained in two other solvent systems (data not shown). Desalted samples of a-aminoadipic acid from the analyzer were then derivatized to the N-trifluoroacetyl n-butyl ester and analyzed by gas liquid chromatography-mass spectrometry. The retention time (1.5 min) of the derivative was the same as the retention time of the derivatized standard on gas liquid chromatography. The mass spectral analysis was performed with a pulsed positive negative ion chemical ionization mass spectrometer which was designed to simultaneously record the positive and negative chemical ionization mass spectra (12). The presence of the trifluoroacetyl butyl ester of a-aminoadipic acid with a molecular weight of 369 was confirmed by the presence of an ion at m/e 370 on the positive ion trace and an ion at m/e 368 on the negative ion trace. Further evidence for the presence of allysine in the major outer membrane proteins of E. coli was obtained by analyzing for a derivative of allysine that is generated by a procedure independent of oxidation or reduction. This procedure recently

3711

has been applied to the analysis of lysyl-derived aldehydes in connective tissue (19, 22), and it is based on the classical Strecker amino acid synthesis. The reaction of an aldehyde with NH3 and CN- results in the formation of an a-amino nitrile, which upon acid hydrolysis, is converted to an a-amino acid. The predicted derivative of allysine is a,c-diaminopimelic acid. A purified preparation of [4-3H]lysine-labeled protein 3a and protein 1 was reacted with NH3 and Na14CN. A single, uncleaved protein band that possessed both 3H and 14C radioactivity was obtained for both proteins. A preparative experiment was then performed with [U-"4C]lysine-labeled protein Sa reacted with nonradioactive CN-. The reacted protein was hydrolyzed and run on the split-stream amino acid analyzer, which used a single column. The radioactive profile from this experiment is given in Fig. 5; in contrast to previous experiments, the radioactive lysine peak is shown. The profile reveals the presence of lysine-derived radioactivity eluting from the analyzer at precisely the calibrated position of a,e-diaminopimelic acid. In contrast to previous derivatization techniques, substantially more of the total lysine radioactivity appears to be present in this derivative (24%). DISCUSSION We have presented evidence for the presence of a-aminoadipic acid b-semialdehyde (allysine) in two outer membrane proteins of E. colt. The implication of this finding is that this compound arose by post-translational oxidative deamination of lysine residues. The possibility remains that the formation of allysine is an artifact arising from the activity of an amino acid oxidase or aminotransferase during preparation of the membranes, but this is unlikely because of the low temperature used during membrane isolation and the short time between when the cells were broken and when the membranes were subjected to denaturing conditions. The presence of allysine was not previously detected for at least two reasons. Conventional amino acid analysis would not detect underivatized allysine, and any ninhydrin-reacting material would be obscured by other amino acids. Second, we omitted 2-mercaptoethanol from the isolation and purification of these proteins since they have been shown to lack cysteine. If 2-mercaptoethanol had been used, it would be expected to have reacted with the aldehyde of the allysine to form a a,e-Diaminopimelic acid

6 35 c

Asp

-0

Glu Ala Met Tyr Lys

Arg

E 30-

E Qa

, 25._

o 200 'D

O -

X

1C

15-

,,10 -

-J

5Origin

Migration

-4

Solvent

front

FIG. 4. Cellulose thin-layer chromatogram of [U-14C]lysinederived radioactivity collected from the amino acid analyzer which eluted at the positions of a-aminoadipic acid and aspartate. (Top) Radioactivity from protein 3a; (Middle) radioactivity from protein 1; (Bottom) authentic a-['4Claminoadipic acid. The solvent system was butanol/acetic acid/water (12:3:5 vol/vol).

10

20

30 40 Fraction

50

60

70

FIG. 5. Split-stream amino acid analysis of hydrolysate of [U14C]lysine-labeled protein 3a that had been reacted with NaCN and

NH3. The calibrated elution positions of several amino acids are given. This experiment was performed on a single column amino acid analyzer so the lysine radioactivity is illustrated.

3712

Biochemistry: Diedrich and Schnaitman

thioacetal. The only report in the literature that suggested the presence of an unusual amino acid in these proteins was the sequencing study of Rosenbusch (23), who was unable to identify the second amino acid from the NH2 terminus. The results of the oxidation study are difficult to interpret quantitatively. If the measured amount of a-aminoadipic acid reflects the true extent of modification of the lysine residues, then there would be only 1 mole of allysine per mole of protein. However, at least three of the CNBr peptides appear to contain allysine. This could be interpreted to mean that the proteins were modified in an apparently random fashion with an average of only one modified residue per polypeptide or that a subpopulation of protein exists which is extensively modified. If the latter is correct, it does not account for the two forms of protein 1, since proteins la and lb possess approximately the same amount of allysine (unpublished observation). A third alternative is that our quantitation is not correct. This would appear to be the case based on the significantly greater amount of allysine implied by the results of the CN- and NH3 reaction in Fig. 5. A fourth alternative is also possible, which is not exclusive of the other interpretations. Since aldehydes are reactive, it is possible that the bulk of the aldehyde is unavailable for derivatization because it had been consumed previously in another reaction. For example, some of the aldehydes may have engaged in Schiff-base formation with available amino groups, and these aldimines may have been biologically reduced to secondary amines. An extension of the fourth interpretation is that the function of the allysine in bacterial membranes may be similar to its crosslinking function in connective tissue. The nature of the chemical crosslinks would not have to be identical to those found in collagen. For example, the allysine could form an aldimine crosslink to the diaminopimelic acid in the murein. In support of the potential crosslink function of allysine in the bacterial membrane, we have recently obtained evidence that a fraction of the major outer membrane proteins remains attached to the murein after the samples are boiled in NaDodSO4. This fraction of protein increases in stationary phase and leads us to suspect that the most extensively modified protein is in this fraction. Thus, in this study, we may have been examining only an intermediate fraction of the protein destined to be bound to the murein. While these studies were in progress, D. Mirelman and R. Siegel (abstract, Israeli Biochemical Society, Isr. J. Med. Sc., in press) made an identical discovery of lysine aldehydes. Their studies indicate that the lysyl-derived aldehydes are involved Schiff-base crosslinking, possibly to free amino groups of the peptidoglycan.

Proc. Natl. Acad. Sci. USA 75 (1978) We thank D. C. Benjamin and R. G. Langdon for assistance with the amino acid analysis, D. F. Hunt for the mass spectral analysis, D. Louth for technical assistance, and D. Mirelman for communicating results prior to publication. Support was by Grant BMS-04973 from the National Science Foundation and Public Health Service Grant GM18006 from the National Institute of General Medical Sciences. D.L.D. was supported by Public Health Service Fellowship A105049 from the National Institute of Allergy and Infectious Diseases. 1. Mulradt, P. F. & Gulecki, J. R. (1975) Eur. J. Biochem. 51, 343-352. 2. Kamio, Y. & Nakaido, H. (1976) Biochemistry 15,2561-2570. 3. Schnaitman, C. A. (1974) J. Bacteriol. 118,442-453. 4. Lugtenberg, B., Meijers, J., Peters, R. & van der Hoek, P. (1975) FEBS Lett. 58, 254-258. 5. Diedrich, D. L., Summers, A. 0. & Schnaitman, C. A. (1977) J. Bacteriol. 131,598-607. 6. Schmitges, C. J. & Henning, U. (1976) Eur. J. Biochem. 63, 47-52. 7. Bassford, P. J., Jr., Diedrich, D. L., Schnaitman, C. A. & Reeves, P. (1977) J. Bacteriol. 131, 608-622. 8. Schnaitman, C. A., Smith, D. & Forn de Salsas, M. (1975) J. Virol. 15, 1121-1130. 9. Miller, E. J., Pinnell, S. R., Martin, G. R. & Schiffman, E. (1967) Biochem. Biophys. Res. Commun. 26, 132-137. 10. Franzblau, C., Sinex, F. M. & Faris, B. (1965) Nature (London) 205,802-803. 11. Roach, D. & Gehrke, C. (1969) J. Chromatog. 44,269-278. 12. Hunt, D. F., Stafford, G. C., Jr., Crow, F. W. & Russell, J. W. (1976) Anal. Chem. 48,2098-2105. 13. Wickner, R. B., Tabor, C. W. & Tabor, H. (1970) J. Biol. Chem.

245,2132-2139. 14. George, D. J. & Phillips, A. T. (1970) J. Biol. Chem. 245,528537. 15. Egan, R. M. & Phillips, A. T. (1977) J. Biol. Chem. 252,57015707. 16. Paz, M. A., Henson, E., Rombauer, R., Abrash, L., Blumenfeld, 0. & Gallop, P. (1970) Biochemistry 9,2123-2127. 17. Gray, W. R. (1967) in Methods in Enzymology, ed. Hirs, C. H. W. (Academic, New York), Vol. 11, pp. 131-151. 18. Lent, R. W., Smith, B., Salcedo, L. L., Faris, B. & Franzblau, C.

(1969) Biochemistry 8,2837-2845.

19. Pereyra, B., Blumenfeld, O., Paz, M., Henson, E. & Gallop, P. (1974) J. Biol. Chem. 249, 2212-2219. 20. Schnaitman, C. A. (1973) Arch. Biochem. Biophys. 157, 541552. 21. Skurray, R. A., Hancock, R. E. W. & Reeves, P. (1974) J. Bac-

teriol. 119,726-735. 22. Paz, M. A., Keith, D. A., Traverso, H. P. & Gallop, P. (1976) Biochemistry 15, 4912-4918. 23. Rosenbusch, J. (1974) J. Biol. Chem. 249, 8019-8029.

Lysyl-derived aldehydes in outer membrane proteins of Escherichia coli.

Proc. Nati. Acad. Sci. USA Vol. 75, No. 8, pp 3708-3712, August 1978 Biochemistry Lysyl-derived aldehydes in outer membrane proteins of Escherichia...
1MB Sizes 0 Downloads 0 Views