JOURNAL OF BACTERIOLOGY, Oct. 1968, p. 1159-1170 Vol. 96, No. 4Copyright ( 1968 American Society for Microbiology Printed in U.S.A.

Lipid Composition of the Electron TransportMembrane of Haemophilus parainfluenzae

DAVID C. WHITE

Biochemistry Department, University of Kentucky Medical Center, Lexington, Kentucky 40506

Received for publication 13 June 1968

The principal lipids associated with the electron transport membrane of Hae-mophilus parainfluenzae are phosphatidylethanolamine (78%), phosphatidyl-monomethylethanolamine (0.4%), phosphatidylglycerol (18%), phosphatidyl-choline (0.4%), phosphatidylserine (0.4%), phosphatidic acid (0.2%), and car-diolipin (3.0%). Phospholipids account for 98.4% of the extractible fatty acids.There are no glycolipids, plasmalogens, alkyl ethers, or lipo amino acid esters inthe membrane lipids. Glycerol phosphate esters derived from the phospholipids bymild alkaline methanolysis were identified by their staining reactions, mobility onpaper and ion-exchange column chromatography, and by the molar glycerol tophosphate ratios. Eleven diacyl phospholipids can be separated by two-dimensionalthin-layer chromatography. Each lipid served as a substrate for phospholipase D,and had a fatty acid to phosphate ratio of 2:1. Each separated diacyl phospholipidwas deacylated and the glycerol phosphate ester was identified by paper chromatog-raphy in four solvent systems. Of the 11 separated phospholipids, 3 were phospha-tidylethanolamines, 2 were phosphatidylserines, and 2 were phosphatidylglycerols.Phosphatidylcholine, cardiolipin, and phosphatidic acid were found at a singlelocation. Phosphatidylmonomethylethanolamine was found with the major phos-phatidylethanolamine. Three distinct classes of phospholipids are separable ac-cording to their relative fatty acid compositions. (i) The trace lipids consist of twophosphatidylethanolamines, two phosphatidylserines, phosphatidylcholine, phos-phatidic acid, and a phosphatidylglycerol. Each lipid represents less than 0.3% of thetotal lipid phosphate. These lipids are characterized by high proportions of theshort (C10 to C14) and long (C19 to C22) fatty acids with practically no palmitoleic acid.(ii) The major phospholipids (93% of the lipid phosphate) are phosphatidylethanol-amine, phosphatidylmonomethylethanolamine, and phosphatidylglycerol. Theselipids contain a low proportion of the short (<C14) and long (>Cl9) fatty acids.Palmitic and palmitoleic acids represent over 80% of the total fatty acids. (iii) Thefatty acid composition of the cardiolipin is intermediate between the other twoclasses. Both palmitoleic and the longer fatty acids represent a significant proportionof the total fatty acid.

Haemophilus parainfluenzae has been shown torequire a membrane-bound electron transportsystem for metabolic activity (24). The mem-brane-bound electron transport system consistsof dehydrogenases, cytochromes, cytochrome oxi-dases, a respiratory quinone, and phospholipids(20, 29, 32). Of the total fatty acids that can beextracted with organic solvents, 98% are asso-ciated with phospholipids (26). However, treat-ment.of the membrane with organic solvents thatextract lipids produces a simultaneous disruptionof electron transport (23). This property of thephospholipids in electron transport prompted anexamination of the lipids in the membrane of H.parainfluenzae. The composition of the mem-

brane-bound electron transport system can beextensively modified during growth in variousenvironments (20, 21, 23, 25). The concentrationof 2-demethyl vitamin K2 has been shown tochange during these environmental modifications(22). If the. phospholipid composition is known,the changes in the phospholipid composition canbe correlated to the modified composition of therespiratory pigments. This could lead, hopefully,to deeper understanding of the formation of themembrane complex.

This study will show that, in addition to thephospholipids expected in a typical gram-negativeaerobe, H. parainfluenzae synthesizes small quan-tities of phospholipids which differ markedly in

1159

on May 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.

fatty acid composition from the major lipids. H.parainfluenzae forms traces of a lipid with prop-erties like phosphatidylcholine.

MATERIALS AND METHODS

Cultures. The medium, cultural conditions, bac-terial strain, and harvesting conditions have beendescribed (20, 24). The dry weight was measured asdescribed previously (24). The lipids were labeled bygrowing the H. parainfluenzae in the presence of 100,uc of H332PO4 per 1.5 liters of media for at least 21generations, unless otherwise stated.

Materials. Phosphatidylglycerol and phosphatidyl-ethanolamine isolated from Escherichia coli weregifts from J. H. Law. Cardiolipin was obtained fromSylvana Chemical Corp., Englewood Cliffs, N.J.,or was isolated from Staphylococcus aureus (27).Phosphatidylserine and phosphatidylcholine wereobtained from the Applied Science Laboratories,Inc., State College, Pa. Labeled glycerolphosphoryl-monomethylethanolamine and glycerolphosphoryl-choline were gifts from R. L. Lester. Carrier-freeH332P04 was supplied in plastic bottles by Tracerlabs,Richmond, Calif. Other materials were of the bestgrade available commercially, or as described pre-viously (23, 24).

Extraction of the lipid. The two methods used toextract the lipids have been described (26). For large-scale preparations, H. parainfluenzae was harvested,washed, and suspended in a mixture of chloroformand methanol (2: 1, v/v) and extracted at roomtemperature for 2 hr with stirring. The extract wasthen filtered through glass wool, and the residue wasreextracted with chloroform-methanol twice. Thecombined extracts were partitioned against salt water(3), and nonlipid components were removed bychromatography with Sephadex (18) from which thephospholipids were recovered quantitatively. Thesecond method of extraction was the Bligh andDyer procedure (2). Various extraction procedureswere used in an attempt to maximize the total lipidphosphate that could be extracted. Exposure of thecells to ultrasonic vibration, boiling the cells inisopropanol before extraction, acidifying the growthmedium before harvesting the bacteria, or boilingin isopropanol after acidification resulted in the re-covery, respectively, of 39.1, 55.4, 57.9, and 70.2,umoles of phospholipid phosphate per g (dry weight)of cells. The Bligh and Dyer and the chloroform-methanol methods (2, 26) resulted in the extractionof 76.6 i 1.1 /Amoles of phospholipid phosphate perg (dry weight) of cells. Boiling in isopropanol andacidification of the medium before harvest is necessaryfor the maximal recovery of the lipids of some bac-teria (27). The proportions of the lipids, determinedby chromatography of the deacylated derivatives onDowex-1, were essentially the same for all modifi-cations of the extraction procedure. Acidifying thechloroform-methanol does not extract additionalphospholipid (26). These extractions leave a residuewith distinctively different fatty acid compositionfrom the fatty acid composition of the total cells (26).

Thin-layer chromatography. The lipids chroma-

tographed over thin-layer plates of silica gel G weredetected with reagents and recovered from the silicagel by methods described previously (27). The solventsused for separating the intact lipids were chloroform,methanol, and 6.7 M ammonium hydroxide(33:18.2:2.5, v/v) in the first dimension, and chloro-form, methanol, acetic acid, and water (45:7:3.15:0.5,v/v) in the second dimension. The lipid was dried bysuccessive evaporations with benzene and absoluteethyl alcohol (4:1, v/v; 33), and not more than 0.5,umole of lipid phosphate was applied to each platefor maximal reproducibility. The silica gel was 0.5mm thick. For separation of larger amounts of thephospholipids, the lipids were fractionated as follows:(i) phosphatidylcholine and neutral lipids (RF value,1.0 to 0.8); (ii) the major and two minor phosphatidyl-ethanolamines and part of the cardiolipin (RF value,0.8 to 0.62); (iii) the major and minor phosphatidyl-glycerols, the phosphatidylserine, and the rest of thecardiolipin (RF value, 0.62 to 0.38); and (iv) thephosphatidic acid and two trace phosphatidyl-ethanolamines (RF value, 0.38 to 0.01). This wasaccomplished by one-dimensional chromatographyof a strip that contained 5 ,umoles of lipid phosphateper plate in a solvent of chloroform, methanol, and6.7 M ammonium hydroxide (33:18.2:2.5, v/v).The fraction that contained phosphatidylcholine waschromatographed on thin-layer plates in chloroformand isooctane (1:1, v/v) to remove the neutral lipid.Each of the fractions from the initial chromatographywas then chromatographed in the two-dimensionalsystem described above. Several plates were used foreach fraction from the preliminary purification. Thecardiolipin and the trace phosphatidylethanolaminesfrom the preliminary fractionation were combinedinto three fractions, since they occupied similar posi-tions on the two-dimensional thin layer plates. Thelipids were then recovered. Those occupying the sameposition were pooled, and about 10% of the recoveredlipid was subjected to mild alkaline methanolysis.The glycerol phosphate derivatives were identified bypaper chromatography as described below. The pooledlipids contained only one glycerol phosphate ester ineach spot.

Mild alkaline methanolysis. A modified procedureof alkaline methanolysis developed by R. L. Lester(unpublished data) was used to deacylate the extractedlipids. The lipid was dissolved in 1 ml of methanoland toluene (1:1, v/v), and an equal volume of freshlyprepared methanolic 0.2 M KOH was added (20jAmoles per ,umole of phosphate). The solution wasincubated at 0 C until the reaction was completed.The solution was then neutralized by adding a suspen-sion in water of Biorex-70 (Bio-Rad Laboratories,Richmond, Calif.) in the acid form (1 meq/10 ,umolesof phosphate). The mixture was added to a columnin a Pasteur pipette that contained an additional 3.3meq of Biorex. The test tube that contained the mix-ture was rinsed twice with 1 ml of methanol and thentwice with 2 ml of water. Each rinse was added to thetop of the Biorex column. The solvents were forcedout of the Pasteur pipette with compressed air. Theeluate was pH 7.0 and it contained less than 1 ,mole

1160 WHITE

on May 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

VOL. 96, 1968 LIPIDS OF H. P

of K+, as measured with a Coleman flame pho-tometer. This method gave the lowest proportion ofa glycerol phosphate in the lipids.

The methanolysis was performed at 0 C to preventthe side reactions which result in cyclic glycerol phos-phate and other secondary products. The time courseof the methanolysis is illustrated in Fig. 1. The re-covery of phosphate in the aqueous phase is 100 i0.5%. As will be established, the products of themethanolysis are glycerol phosphate esters in theaqueous phase, and methyl esters of the fatty acidsin the organic phase. In early experiments involvingcolumn chromatography of the glycerol phosphateesters, the methanolysis was stopped by neutraliza-tion with acetic acid. Under these conditions, between1.5 and 2.0% of the lipid phosphate was recoveredas a-glycerol phosphate. The use of Dowex-50 toneutralize the base has been suggested by Benson (1).With these anhydrous conditions, neutralization withDowex-50 resulted in large proportions of a-glycerolphosphate.The following abbreviations are used to represent

the glycerol phosphate esters derived from the phos-pholipids by deacylation: GPC, glycerolphosphoryl-choline derived from phosphatidylcholine; GPE,glycerolphosphorylethanolamine derived from phos-phatidylethanolamine; GPMME, glycerolphosphoryl-monomethylethanolamine derived from phospha-tidylmonomethylethanolamine; GPDME, glycerol-

5.0

C,)

O 4.0

0

C:) 3.00

a.

X, 20w-10

1.0

_k WATE R

to CHCI 3

1 2 3HOURS OF INCUBATION AT OC

FIG. 1. Time course of the mild alkaline methanoly-sis of the lipids of H. parainfluenzae. The lipid wasdissolved in I ml of methanol and toluene (1: 1).Freshly prepared 0.2 M methanolic KOH (I ml) was

added and the mixture was incubated at 0 C. At thetimes indicated, portions were removed, sufficientI M acetic acid was added to bring the mixture to pH7.0, and equal volumes of both chloroform and waterwere added. The solution was mixed with a vortexmixer for 5 min. The aqueous phase was removed andcombined with a second aqueous phase that also hadbeen partitioned against the chloroform. The volumeswere reduced to dryness in a stream of nitrogen; thelipids were digested and analyzed for phosphate.

'ARAINFLUENZAE 1161

phosphoryldimethylethanolamine derived fromphosphatidyldimethylethanolamine; GPG, glycerol-phosphorylglycerol derived from phosphatidylgly-cerol; GPS, glycerolphosphorylserine derived fromphosphatidylserine; aGP, L-a-glycerolphosphatederived from phosphatidic acid; GPGPG, diglycer-olphosphorylglycerol derived from cardiolipin.

Column chromatography. The gradient fractiona-tion of deacylated phospholipids on Dowex-1 columnswith ammonium formate-sodium borate has beendescribed (27). Components eluting in the early frac-tions of the column can be better resolved by theuse of 0.02 M ammonium formate at pH 9.0 (R. L.Lester, unpublished data). This procedure was usedto detect GPC. The individual glycerol phosphateesters separated by preparative columns were pooledand desalted, as will be described in the Results.

Paper chromatography. Glycerol phosphate estersderived from the phospholipids by deacylation wereseparated by paper chromatography by the use ofsolvents and detection methods described previously(27). GPE, GPMME, and GPDME were separatedby paper chromatography in a solvent of 3.8 mMethylenediaminetetraacetic acid and 0.7 M ammoniumbicarbonate in 90 mm ammonium hydroxide con-taining 67% ethyl alcohol. The components weredetected by radioautography or an amine reagent(9). Amino cellulose paper (Whatman AE-81) wasfirst treated with 3M formic acid and dried. Thesamples were then chromatographed over a solventsystem of 0.4% pyridine in 3 M formic acid in thefirst dimension, and in Wawszkiewicz solvent I (17)diluted 3 to 7 with 95% ethyl alcohol containing0.26 M ammonia in the second dimension. Up to 2.5iAmoles of lipid phosphate can be chromatographedwithout streaking. Lipids can be localized by radio-autography or by treatment of the papers withperiodate followed by o-tolidine (27). These methodswere developed by R. L. Lester (unpublished data).

Gas-liquid chromatography. Gas-liquid chroma-tography for the methyl esters of the fatty acids wasperformed as previously described (26). Fatty acidswere designated with a subscript for the length ofthe carbon chain, and with "Br" for iso and anti-isobranched acid esters, "C" for cyclopropane, and 1following a colon (:1) for monounsaturated (26).The fatty acid methyl esters were prepared andidentified as described (26). Gas-liquid chromatog-raphy of the lipid acid hydrolysate was used to identifythe ethanolamine-containing lipids (10).

Analysis of the lipids. Analysis of the lipids forglycerol, fatty acids, phosphate, carbohydrate, andamino nitrogen were performed as described pre-viously (27).

Treatment with phospholipase D. Between 0.1 and0.01 jAmole of lipid phosphate was dissolved in 1 mlof ether in a test tube, and 0.42 ml of water that con-tained 20 ,umoles of acetate buffer (pH 5.6), 40 jAmolesof CaCl2, and 1.6 mg of cabbage phospholipase D(Calbiochem, Los Angeles, Calif.) was added. Thereaction mixture was agitated vigorously and in-cubated at 25 C for 4 hr. The ether was removed witha stream of nitrogen, and 1.85 ml of water, 2.5 ml of

on May 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.

chloroform, and 2.5 ml of methanol were added.After thorough mixing, the lower phase was recovered,and the lipid was deacylated. The water-soluble glyc-erol phosphate esters were identified by paper chro-matography with a water-saturated phenol solvent.

Measurement of radioactivity. Samples were as-sayed for radioactivity in a model 2311 scintillationspectrometer (Packard Instrument Co., Inc., Down-ers Grove, Ill.). Deacylated phospholipids werecounted on paper discs (1.5 to 2 cm in diameter) in ascintillation fluid of 9.28 mm 2, 5[2 (5-tertbutylbenzoxazol)]-thiophene (BBOT) in toluene. Sinceall counts were made on similar paper discs, the in-fluence of quenching was minimized. The efficiencyof counting for 32p was 90.6%. No radioactivity couldbe detected in the vial after removal of the paper disc.Water-soluble materials in solution were countedin a scintillator consisting of 1.236 M naphthalene and6.99 mm BBOT in dioxane. Up to 1 ml of aqueoussolution can be put with 10 ml of counting fluid. Theefficiency of this scintillator for 3^P was 100%.

Radioautography. Dried papers or Saran-wrappedthin-layer plates were placed on Kodak no-screenX-ray film for 12 to 72 hr in film holders. The filmwas then developed, and the dark spots correspond-ing to radioactive components were used to find thecomponents on the chromatogram.

RESuLTSOver 98% of the fatty acid extractable with

organic solvents from H. parainfluenzae are boundto phospholipids (26). Consequently, phospho-lipids of this organism were analyzed. Mildalkaline methanolysis, of diacyl phospholipidsconverts the lipids into water-soluble glycerolphosphate esters and methyl esters of the fattyacids that are soluble in organic solvents. Theinitial analysis of the lipids of H. parainfluenzaeinvolved the separation and identification of theglycerol phosphate esters. Lipid phosphate canbe quantitatively recovered in the aqueous phaseafter mild alkaline methanolysis, as illustrated inFig. 1. This fact precludes the presence of alkylethers and plasmalogens in the lipids of H. parain-fluenzae. Theseether derivatives are not hydrolyzedunder the conditions of mild alkaline methanolysis(19).Analysis of a lipid sample containing 27 ,umoles

of lipid phosphate for carbohydrate with theanthrone reagent (27) indicates that less than 0.1,umole of carbohydrate was present in the lipid.After thin-layer chromatography of 5 ,umoles oflipid phosphate, no reactivity with diphenylaminereagent (16) was seen. The diphenylamine reagentused in this manner is sensitive to 0.05 ,umole ofmonoglycosyldiglyceride per cm2. Within experi-mental error, all of the fatty acids from the com-plex lipids recovered from a silicic acid columnwith methanol are accounted for by the phospho-lipids (26). Lipids isolated from H. parainfluenzae

grown with "4C-glucose or '4C-glycerol and 32p,and separated by chromatography, give no evi-dence for complex lipids other than phospho-lipids.

Separation of the glycerol phosphate esters.Water-soluble glycerol phosphate esters of thephospholipids were separated by chromatographyon Dowex-1 with the gradient described pre-viously (27). Deacylated phospholipids derivedfrom bacteria grown in the presence of 32P areillustrated in Fig. 2. Repeated analysis indicatesthat the proportions of deacylated phospholipidsare reproducible to i 1% for GPE and GPG andto a0.2% for the other lipids. A tentative iden-tification of the lipids can be made by comparisonof the volume necessary to elute a volume ofspecific glycerol phosphate ester. The elutionvolumes are reproducible.

Several column separations similar to thatshown in Fig. 2 yielded a total of 394 umoles oflipid phosphate distributed among the six tenta-tively identified glycerol phosphate esters. Eachglycerol phosphate derivative was pooled anddesalted on a second Dowex-1 column. Boratewas removed by washing the column with severalvolumes of 0.05 M formic acid. The glycerol phos-phate esters were then eluted with 0.03 M am-monium formate. The ammonium ion was re-moved by passing the eluate through a third

a =GPC GPE

cGPG aGP GPS

6 mGPGPG GPGP(?)

FIG. 2. Chromatographic analysis of the deacylatedphospholipids from H. parainfluenzae grown in thepresence of 32P. Lipids were eluted from a Bio-RadAG-I (Dowex-1) column in the formate form with thegradient described previously (27). The gradient waspumped through the column at a rate of 0.4 ml/min,and 2.64-ml fractions were collected. Each fractionlwas assayed for 32p by removal of 0.5 ml, which wasadded to the dioxane scintillation mixture and theradioactivity determined. The remainder of the frac-tion was taken to dryness at 150 C, digested, and thephosphate determined. Recovery of 32p added to thecolumn was 97%. The RF values of known deacylatedlipids are illustrated at the top of the figure.

1162 WHITE

on May 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

LIPIDS OF H. PARAINFLUENZAE

Dowex-50 (acid form) column. The eluate wascollected in ice and lyophilized as rapidly aspossible. Recovery of the lipid phosphate afterdesalting for each deacylated lipid was GPC,71%; GPE, 99%; GPG, 78%; GPS, 49%; aGP,62%; and GPGPG, 74%.

After desalting, each component was subjectedto paper chromatography in four solvent systems.The results of this chromatography are illustratedin Table 1. Each of the deacylated phospholipidsgave the expected reactions for phosphate, aminonitrogen, choline, and vicinyl glycols.

Analysis of the deacylated phospholipids. Themolar glycerol to phosphate ratios of the deacyl-ated phospholipids used in the paper chromatog-raphy described above were GPE, 0.92:1.00;GPG, 1.98:1.00; aGP, 0.98:1.00; GPS, 1.05:1.00; and GPGPG, 1.56:1.00. Insufficient mate-rial was available for the analysis of GPC.

Enzymatic assaysfor aGP. The enzymatic assayof Wichert (30) for L-a-glycerol phosphate wasused with the fraction that had the chromato-graphic and analytical characteristics of aGP. In

TABLE 1. Paper chromatography of the deacylatedphospholipids isolatedby column chromatography

from Haemophilus parainfluenzae

RF valuesLipida

Solvent 1 Solvent 2 Solvent 3 Solvent 4

GPC 0.91 0.35 0.97 0.55GPE 0.59 0.17 0.97 0.48GPG 0.47 0.24 0.38 0.48aGP 0.23 0.33 0.32 0.05GPS 0.33 0.15 0.62 0.23GPGPG 0.10 0.08 0.08 0.12

a Deacylated phospholipids were separated bycolumn chromatography similar to that illustratedin Fig. 2. Each component was pooled and de-salted as described in the text. Solvent 1 was water-saturated phenol adjusted to pH 5.5 (27); solvent2 was butan-l-ol, acetic acid, and water (5:3:1,v/v) (27). These solvents were used with Schleicherand Schuell (no. 589) acid-washed paper. Solvent3 was 3 M formic acid containing 0.4% pyridine(v/v); solvent 4 was Wawszkiewicz solvent (17)diluted 3 to 7 with 95% ethyl alcohol containing0.26 M ammonia. The paper was Whatman AE-81amino cellulose paper treated with 3 M formicacid. Each deacylated lipid was both comparedand co-chromatographed with the correspondingdeacylated authentic lipid by one-dimensionalascending chromatography. The deacylated lipidsgave the expected reactions for amino-nitrogenwith ninhydrin, phosphate with acid molybdate,and vicinyl glycol with periodate and o-tolidineusing techniques described previously (27). TheGPC reacted with Dragendorf's reagent (16).

the sample of aGP, 96.5% of the aGP deter-mined by phosphate analysis can be accounted forby the enzymatic assay.

Analysis of the nitrogen-containing phospho-lipids. The GPE fraction from the preparativeDowex-1 columns was hydrolyzed in 3 M HCI for3 hr at 100 C. The hydrolysate was neutralizedwith solid NaCO3, extracted with tertiary butanol,and subjected to gas-liquid chromatography asdescribed (10). The molar ratio of ethanolamineto glycerol with the ethanolamine determinedfrom gas-liquid chromatography was 1.02:1.00.The ethanolamine had a relative retention volumeof 1.8 compared to an internal standard of cyclo-hexylamine.

In a total 'lipid mixture that was deacylated,acid-hydrolyzed, and analyzed by gas-liquid chro-matography, monomethylethanolamine was de-termined to be 0.35% of the total lipid phosphate(10). The monomethylethanolamine isolated fromthe lipid had retention volume identical to that ofan authentic sample of monomethylethanola-mine.A sample of Haemophilus lipid was partitioned

twice against 4 M KCI atpH 2.0, chromatographedon a preparative thin-layer plate, eluted from thesilica gel, and treated with Sephadex (19). Thisprocedure was necessary to remove a peptidecontaminant from the lipid. No amino acidmethyl esters could be detected in the aqueousphase after mild alkaline methanolysis. Therefore,o-lipoyl esters of amino acids found in gram-positive bacteria by Macfarlane (11) were notpresent in Haemophilus. After removal of thepeptide contaminant, the lipid was deacylatedand hydrolyzed in 2 M HCI under nitrogen for 3hr at 115 C. The amino nitrogen-containing com-ponents in the hydrolysate were separated and as-sayed with a Technicon amino acid analyzer. Twomajor components were detected. One had anRF value identical to that of ethanolamine, andthe second had an RF value identical to that ofserine. The ratio of GPE to GPS, calculated fromthe phosphate after separation of the glycerolphosphate esters bychromatography on Dowex-1,equaled the ratio of ethanolamine to serine meas-ured after acid hydrolysis of the lipid and separa-tion with the amino acid analyzer. About 99.2%of the ethanolamine of GPE was recovered, andabout 94% of the GPS was recovered.. The re-maining 0.03% of the total amino nitrogen-reac-ting material was glycine, which may have comefrom the serine during hydrolysis.

Presence of GPC. Phosphatidylcholine hasbeen reported to be restricted to bacteria withextensive internal membrane systems (4). Haemo-philus does not show extensive internal mem-branes, yet it contains traces of a lipid which

1163VOL. 96, 1968

on May 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.

yields GPC upon deacylation. Glycerol phosphateesters containing 2P were eluted from a Dowex-1column with 20 mm ammonium formate (pH 9.0)(Fig. 3). The fractions containing the glycerolphosphate ester with the elution volume of GPCwere pooled and the ammonium formate wasremoved by evaporation in vacuo. The samplewas co-chromatographed with authentic GPC andseparated from GPE (Fig. 4). The sample gavereactions characteristic of choline-containinglipids with phosphomolybdate after ninhydrin(15), Dragendorf's reagent (16), and hydratedcobalt-amine complex (9). H. parainfluenzae in-corporated 32p from the growth medium into alipid characterized as containing a quaternaryamine that has the chromatographic propertiesof GPC after deacylation. As yet, it has not beenpossible to isolate and identify choline as the solenitrogen-containing component of acid-hydro-lized GPC. Consequently, the designation of thisdeacylated lipid as GPC must be consideredtentative.

Separation ofthe lipids by thin-layer chromatog-raphy. Phospholipids isolated from H. parainflu-enzae can be separated by two-dimensional chro-matography with silica gel G thin-layer plates. Aradioautogram of the chromatographed lipids isillustrated in Fig. 5. Each of the separated lipids

GPC MM-GPE GPEco R-

0.05w

uw> 0.04

x

2 0.03.a.C-

wJ 0.02m

L .oJ 0.01

w

0

2 L

C.0.5 .

w

ZND0.4

LiX n

I -i2

02O

x 0.1x °00oX OxxX o OoXo0

DM-GPE

o*0;

IX X

x VI le, RS,IX -

100

80 _-OS

60 F'o

40x

2a-

20

ZL 10 20 30 40 50 SO

TUBE NUMBER

FIG. 3. Chromatographic analysis of the ethanola-mine deacylated phospholipids of H. parainfluenzae.A column like that described in Fig. 2 was loaded withdeacylated phospholipids. The deacylated phospho-lipids were then eluted with 0.02 ammonium formate,(pH 9.0) and 2.40-mlfractions were collected. Thefrac-tions were assayed as in Fig. 2, and 83.2% ofthe addedphosphate was recovered. The GPG contaminatedwith aGP can be eluted after about 170 ml ofammoniumformate has been used. The GPS and GPGPG canbe eluted only if the pH of the eluting buffer is raised.The elution patterns of authentic deacylated lipids areillustrated at the top of the figure.

r SOLVENT FRONTS %

t-..-.--' GPC IC]

0a I0 Il. Iz IZ Iz

z

as07)

XZ. WAWSZKIEWICZ

IIIII

0 0 11

FIG. 4. Two-dimensional chromatography of GPC-32p on amino cellulose paper. The presumed GPCfraction of H. parainfluenzae grown in the presenceof 82P was collectedfrom columns like those illustratedin Fig. 3. The ammonium formate was removed bysuccessive evaporations in a vacuum oven; the materialwas combined with unlabeled GPE and GPC and sub-jected to chromatography on amino cellulose paperas described in Table 1. After chromatography, thepaper was placed on Kodak no-screen X-ray film for2 days. The paper was then treated with periodatefollowed by o-tolidine (27) to detect vicinyl glycols,and the positions of the unlabeled GPE and GPC weremarked. The radioautogram was then placed over atracing of the chromatogram and a picture was taken.

was eluted from the silica gel as previously de-scribed (27), and was subjected to mild alkalinemethanolysis. The resulting water-soluble glycerolphosphate esters were identified by paper chro-matography with water-saturated phenol solventand by two-dimensional chromatography onamino cellulose paper. In addition, paper chro-matography in the Wawszkiewicz solvent, modi-fied to contain ammonium carbonate, was used toseparate GPE, GPMME, and GPDME. Sincethe specific radioactivity of the three major lipidswas equal, the radioactivity was used to calculatethe proportions of phosphate in each fraction(Table 2). The analyzed proportions of the majorclasses of phospholipids, after separation by thin-layer chromatography, were the same with glyc-erol phosphate esters determined after deacyla-tion and paper chromatography (Tables 2 and 3).Three classes of phosphatidylethanolamines, as

well as two classes each of phosphatidylglyceroland phosphatidylserine, are separable by thin-layer chromatography. The phosphatidylmono-

1164 WHITE

I

on May 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

LIPIDS OF H. PARAINFLUENZAE

methylethanolamine is found with the major frac-tion of phosphatidylethanolamine. The lipid usedfor the analysis shown in Table 2 and in Fig. 5contained a somewhat larger proportion of phos-phate and fatty acids at the origin. Usually theorigin accounts for less than 0.01 % of the lipidphosphate.One obvious reason for the separable types

of phosphatidylethanolamines, phosphatidyl-serines, and phosphatidylglycerols would be thatthe minor components were lyso-phospholipids.Determination of the fatty acid to phosphateratio by three methods rules out lyso-phospho-lipids.The mild alkaline methanolysis results in the

quantitative liberation of fatty acid methyl esters.With C20:1 methyl ester as an internal standard,a summation of the peak areas of all responses ofthe fatty esters detected after gas-liquid chroma-

S C.> ';ENT A

@.1 - .^.#..

CHC3 rH30H4 E, 7

.9 I'_ D.... v ,to

CH 3COOH

-1~ 5

FIG. 5. Thin-layer chromatography of intact phos-pholipids of H. parainfluenzae. Silica gel G thin-layerplates were prepared (27), and less than 0.5 j,mole oflipid was spotted in the lower left hand corner. Theplate was then chromatographed in two dimensionswith the solvents indicated in the figure. The plate wasdried in the vacuum oven, wrapped in Saran wrap, andplaced in contact with X-ray film. After 24 hr, thefilm was developed and the lipids were recovered fromthe plate (27). Radioactivity was determined afterdeacylation by spotting the glycerol derivatives on

paper discs and counting the paper discs in the toluenescintillation fluid. The proportion of the total lipidphosphate in each lipid is given in Table 2. The re-covery of 32Pfrom the plate was 97%.

tography indicates a fatty acid to phosphateratio of 2:1 (Table 2).Cabbage chloroplast phospholipase D, called

phospholipase C by Kates (7), is specific fordiacyl phospholipids. Incubation of the phospho-lipids with phospholipase D resulted in produc-tion of phosphatidic acid (measured as aGP inTable 4) from all the lipids but cardiolipin.Authentic cardiolipin from S. aureus was notattacked by phospholipase D.

Collection of larger amounts of the trace lipidsallowed colorimetric analysis of the fatty acidsand phosphate. The lipids were first fractionatedby one-dimensional thin-layer chromatography asdescribed in Materials and Methods. Then eachfraction was chromatographed in the usual two-dimensional system. Spots corresponding to thesame position after two-dimensional chromatog-raphy were pooled and 10% of the sample wasused to confirm the identification by paper chro-matography of the deacylated phospholipid. Onlyone glycerol phosphate ester was found in eachspot. Spots 1 and 3 in Fig. 5 corresponded to asingle phosphatidylserine, and the origin con-tained less fatty acid than was found in Fig. 5.The remainder of the lipid (90%) was saponifiedin 3 M KOH in 47% ethyl alcohol at 100 C for 4hr. The saponification mixture was acidified topH 2.0 with HCI, andthe fatty acids were extractedwith three portions of petroleum ether andanalyzed by the Novak procedure (13). Theaqueous portion of the saponification mixture wasacidified further and the phosphate was deter-mined colorimetrically (27). Both the phosphateand the fatty acid assays are accurate to 0.01,imole per sample. At least 0.02 ,mole of phos-phate was present in each sample. The fatty acidto phosphate ratio of the phospholipids isolatedby two-dimensional thin-layer chromatographywas 2:1 (Table 5).

Evidence from the following experimentsstrongly suggests that hydrolytic breakdown isnot the source of the minor lipids. The minorlipid components apparently are not generatedby the thin-layer chromatography. A total phos-pholipid sample was applied to two thin-layerplates. One was subjected immediately to chroma-tography and the other was kept at room temper-ature for 72 hr before being placed in thechromatography solvents. Radioautograms fromthe two plates were essentially identical. If aslightly overloaded plate (0.7 Amole of lipidphosphate) is chromatographed and the majorlipid (spot 7) is eluted and chromatographedrepeatedly, there is progressive purification of thephosphatidyl ethanolamine. After the secondchromatography, traces of GPGPG, GPG, aGP,and GPC in the appropriate positions can be

1165VOL. 96, 1968

T. I

0vT i:,.,z 4f)

I t"Qpl-9D

.r00.

xTQ.k

on May 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

TABLE 2. Separation of the phospholipids of Haemophilus parainfluenzae by thini-layer chromatography

Spot,

012345678910

Counts/min

457711

2,0733,170664495

1,145886,500169,30056,0003,142

Percentage ofphospiiate

0.040.060.190.290.060.040.1077.6015.525.770.29

Components

GPC

0.70.10.567.6

GPE

50.5

100100

98.40.41.3

GPMME

0.7

GPG

39.3

3.4

100

98.72.817.4

aGP

94.22.0

0.10.53.0

GPS

100

98.0

GPGPG

10.2

2.4

0.10.192.4

Ratio offatty acidtO P04

5.6:12.1:12.0:11.6:11.9:12.3:11.6:11.9:11.8:11.7:12.2:1

a Spots were removed from four plates like those illustrated in Fig. 5 and deacylated by mild alkalinemethanolysis (27). The water phase from the methanolysis was subjected to chromatography in foursolvent systems as described in the text. The glycerol phosphate esters were detected by radioautographyand assayed by counting the paper disc in the toluene scintillation fluid. Phosphate was then determinedafter digestion of the paper in perchloric acid. The specific activity of the deacylated lipid from spot7 was 366,300 counts/min per jAmole of P; of spot 8, 349,800 counts/min per ,mole of P; and of spot 9,311,500 counts/min per jAmole of P. These were averaged and used to calculate the proportion of phos-phate in the minor lipids from the radioactivity. The organic phase of the mild alkaline methanolysiswas used to determine the fatty acid methyl ester content. An internal standard of C20:1 (1 nmole/;liter) was added and the methyl esters were subjected to gas-liquid chromatography. The areas of theresponses were calculated and compared to the internal standard. Remethylation (26) established thatall the fatty acids released from the methanolysis were methyl esters.

TABLE 3. Comparison of the proportions of the totallipids of Haemophilus parainfluenzae determinedby thin-layer chromatography of the intact

lipids and by chromatography of the deacy-lated lipids on amino cellulose paper

Thin-layer PaperLipida chromatography chromatography

(intact lipids) (deacylated lipids)

GPE 79.7 (0.2-GPC) 78.6GPG 17.5 18.2GPGPG 2.1 2.3caGP 0.3 0.4GPS 0.3 0.5

A total of 1.899 ;umoles of lipid phosphate wassubjected to thin-layer chromatography (Fig. 5)on four plates. The lipids were detected by radio-autography and recovered, deacylated, and iden-tified by paper chromatography (Table 1). Thetotal phosphate in each lipid was then calculated.The recovery of phosphate for the entire pro-cedure was 89.2%. A second portion of the lipidwas deacylated and separated with amino cellu-lose paper. The lipids were located by radioau-tography and cut out. The radioactivity was de-termined and the papers were digested for phos-phate analysis. The recovery of phosphate was101%.

demonstrated. The lipids were identified by paperchromatography after deacylation by mild alka-line methanolysis. The third chromatography ofthe material in spot 7 showed 99.35% as GPEwith mobility like spot 7, 10.24% as GPE likespot 4, 0.20% as GPE like spot 5, and 0.17% asGPC like spot 10. The fourth chromatogram ofthe material in spot 7 was essentially 100% GPEwith mobility like spot 7 with no minor lipidcomponents detectable. With thin-layer platesthat are not overloaded (<0.5 ;4mole of lipidphosphate per plate), it is possible to isolate spot7 free of traces of other lipids after a singlechromatography (Table 5). The minor lipids arenot generated by the separation procedures

Apparently the minor lipid classes exist in H.parainfluenzae, and their separability can beexplained by the differences in proportions of thefatty acids. The relative proportion of each fattyacid in each type of phospholipid separated bythin-layer chromatography is given in Table 6.There appear to be three classes of lipids. Thetrace lipids contain less than 0.3% of the totallipid phosphate; a high proportion (10 to 50%)of Cis, C20, and C22; relatively less C16 and far lessC16:1 (0.6 to 3%); more C14: I than C14; and higherproportions of either C12, C12: , C13, or C14: 1. The

1166 WHITE J. BACTERIOL.

on May 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

LIPIDS OF H. PARAINFLUENZAE

TABLE 4. Hydrolysis ofHaemophilis phos-pholipids by phospholipase Da

Percentage of aGPSpot Component after enzymatic

hydrolysis

0 GPE, GPG 672 aGP 1053 GPS 904 GPE 1005 GPE 1106 GPG 727 GPE 328 GPG 709 GPGPG 010 GPC 22

a Phospholipids labeled with 32p were sepa-rated on eight thin-layer plates (Fig. 5). The lipidswere located by radioautography and recovered.One-third of the sample was deacylated and thewater-soluble glycerol derivatives were identifiedby paper chromatography with a water-saturatedphenol solvent. The remainder of the lipid wastreated with cabbage phospholipase D in acetatebuffer in the presence of ether (7) for 4 hr at roomtemperature. The lipid was then deacylated andsubjected to paper chromatography. The deacy-lated phospholipids were detected by radio-autography, cut out, and the radioactivity wasdetermined in the scintillation spectrometer. Thematerial in spot 7 amounted to 0.16 jsmole of lipidphosphate. When this material was treated withthe enzyme at 100-fold dilution, the hydrolysiswas complete in 4 hr.

major lipids are phosphatidyl ethanolamine (spot7) and phosphatidyl glycerol (spot 8). These twolipids account for 93.12% of the total lipid phos-phate. The C16 and C16:1 comprise 80% of thetotal fatty acids, and the short and very long fattyacids comprise less than 5% of the total fattyacids. The cardiolipin (spot 9) seems from itsfatty acid composition to be a mixture of the twolipid classes. Cardiolipin has high proportions ofC16 and C16 1 as well as high proportions of thelong fatty acids.

DIscussIoN

The membrane of H. parainfluenzae containsdiacyl phospholipids as the major components ofthe extractable lipids. These lipids are present atconcentrations of 60 to 70 ,umoles of phosphateper g of bacterial dry weight. No lipo amino acids,plasmalogens, alkyl ethers, glycolipids or lyso-phospholipids are present. A trace of phos-phatidylmonomethylethanolamine and a lipidvery much like phosphatidylcholine are the major

TABLE 5. Fatty acid to phosphate ratios ofphospholipids of Haemophilus parain-

fluenzae separated by thin-layerchromatographya

Spot Lipid Ratio of fatty acidto phosphate

0 Mixed 1.47:1.001 + 3 GPS 1.69:1.002 aGP 1.99:1.004 GPE 1.99:1.005 GPE 2.07:1.006 GPG 2.02:1.007 GPE 1.98:1.008 GPG 2.03:1.009 GPGPG 1.83:1.0010 GPC 1.84:1.00

Lipids containing 32p were extracted andfractionated by thin-layer chromatography intofour fractions in the chloroform-methanol-6.7M ammonium hydroxide solvent, as described inthe text. Each fraction was then divided intoseveral portions and chromatographed on thin-layer plates in the two-dimensional system illus-trated in Fig. 5. Lipids with similar positions inthis chromatography were pooled and a portionofeach was deacylated by mild alkaline methanoly-sis. The deacylated derivatives were then identi-fied by paper chromatography with the phenol-water solvent (Table 1). Each pooled fraction gaverise to a single type of glycerylphosphoryl deriva-tive. The remaining lipid from the thin-layerplates was saponified; the fatty acids were ex-tracted; the residue was digested; and the fattyacid to phosphate ratio was determined by colori-metric analysis as described in the text. Spots werenumbered as in Fig. 5 according to their positionafter the two-dimensional chromatography. Thedesignation was confirmed by identification of thedeacylated lipid.

differences between the lipids of H. parainfluenzaeand E. coli (5).The striking feature of the phospholipids of H.

parainfluenzae is that 11 phospholipids can beseparated into three classes of lipids that differin fatty acid composition. The trace lipids, whichtogether comprise less than 0.3% of the totallipid phosphate, have large proportions of thelong and short fatty acids with very little C16: 1.In the major lipids, which comprise 93% of thelipid phosphate, 80% of the fatty acids are C16:1and C16. The short and long fatty acids comprisea small proportion of the fatty acids of the majorlipids. The cardiolipin appears to have features ofboth classes of lipids in having both C16:1 and theshort-chain fatty acids in significant proportions.The proportions of phosphate among the 11phospholipids are quite reproducible and the

VOL. 96, 1968 1167

on May 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

TABLE 6. Fatty acid composition of the phospholipids of Haemophilus parainfluenzae separated by thin-layer chromatography

Fatty acidamethyl esters

12:11213:113:Br1313 :C14:114: Br1414:C15:115: Br1515:C16:116: Br1616:C17:117:Br1717:C18:118:Br1818: C19:119: Br1919:C20:Br2020: C21:12122

Percentage of the total area of the recorder response foreach fatty acid ester in the 11 separated phospholipids

0 1 2 3 4 5 6 7 8 9 lo

0.540.369.800.360.361.814.360.734.360.732.720.361.45

8.171.09

25.41

13.610.910.361.450.73

0.91

9.26

9.26

0.070.75

0.130.203.52

1.630.291.14

1.820.262.930.49

26.340.783.914.692.281.950.52

8.99

13.25

0.46

4.07

19.54

0.220.3114.840.110.393.931.181.230.220.224.211.180.340.092.670.3418.380.172.471.260.84

2.360.224.52

0.56

20.74

0.453.250.040.160.120.028.440.452.280.020.180.890.090.092.90

17.980.272.282.610.86

0.430.094.35

0.6116.16

0.270.33

32.19

10.10

0.320.671.09

0.130.18

22.73

1.150.150.370.040.530.071.320.047.740.180.950.700.44

1.060.1233.04

3.64

3.97

19.38

0.400.632.77

0.230.035.040.131.580.260.461.290.070.152.90

13.850.262.243.431.19

5.34

0.170.1611.180.110.030.122.950.060.930.090.92

0.560.121.060.238.771.01

30.281.632.172.420.620.195.81

9.91

0.22

0.652.80

49.46 15.50

0.030.380.050.020.010.010.56

9.850.010.090.460.010.0141.32

39.930.130.070.040.090.421.560.052.37

0.010.570.07

0.05

1.12

0.030.260.09

0.020.010.55

1.620.030.140.010.350.0138.12

41.850.140.310.230.320.302.290.053.60

1.52

0.360.06

4.82

2.96

0.330.190.96

0.040.032.13

1.910.022.760.570.060.06

22.93

31.970.120.891.080.400.623.17

4.70

5.989.74

1.621.21

6.56

0.270.10

0.100.080.880.061.740.141.060.661.061.060.670.26

44.230.432.812.031.600.032.200.826.06

0.674.883.84

0.200.581.272.50

18.86

a Fatty acids methyl esters were taken from the organic phase of the mild alkaline methanolysis of thelipids isolated after chromatography illustrated in Table 2. The volume was reduced in a stream ofnitrogen, and the methyl esters were subjected to gas-liquid chromatography as described previously(26). The fatty acid methyl esters were identified by their retention times on polar and nonpolar columnsbefore and after hydrogenation (26). The numbers represent the percentage of the total areas of the re-corder response for each fatty acid ester taken after separation with a SE-30 column. The molar responseof the hydrogen flame detector to the internal standard C20:, methyl ester did not differ significantlyfrom that expected for the saturated and monounsaturated methyl esters of C12, C14, C16, C,8, and C20determined with known mixtures. Fatty acid methyl esters are designated as the number of carbonatoms, with ":1" for monounsaturated, ":Br" for branched, and ":C" for cyclopropane.

trace components are neither lyso-phospholipids ences is that precursor lipids such as phosphatidyl-nor are they generated during thin-layer chroma- serine, which is decarboxylated to phosphatidyl-tography itself. The possibility that the trace lipids ethanolamine; phosphatidylglycerol, which is thecould represent different salt forms still does not precursor of cardiolipin; and phosphatidic acid,change the fact that the fatty acid composition of which is the precursor of all the phospholipids,the three lipid classes is different. have different fatty acid compositions from theThe interesting feature of the fatty acid differ- end product lipids. The synthesis of phospho-

1168 WHITE J. BACTERIOL.

on May 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

LIPIDS OF H. PARAINFLUENZAE

lipids, as illustrated with E. coli (6), does not con-sider different relative fatty acid compositionsbetween precursor and end product lipids. Pre-liminary experiments indicate that phosphatidyl-ethanolamine is synthesized by decarboxylationof phosphatidylserine in H. parainfluenzae, yetthe fatty acid composition of the major phos-phatidylethanolamine is different from the phos-phatidylserine. Either there is remarkable fattyacid specificity in the enzymes that form the majorphospholipids and rapid turnover of the traceprecursor lipid with the acid composition ofthe major lipid class, or there is modificationof the fatty acid composition after synthesis ofthe lipids. Preliminary experiments with H. para-influenzae have not demonstrated a very rapidturnover or incorporation of radioactivity intophosphatidic acid or phosphatidylserine. Itseems clear that the fatty acid compositions aremodified after the synthesis of the phospholipids.It could be that the trace fraction of phosphatidyl-ethanolamine and phosphatidylglycerol representslipid where the fatty acid composition has notbeen modified significantly from that of the pre-cursor phosphatidic acid.

Recently, the activity of fatty acyl transferasehas been suggested by the differences in the fattyacid composition of the phosphatidylglycerol,lysyl-phosphatidylglycerol, cardiolipin, mono-and diglucosyldiglyceride in S. aureus (28). Thistype of transferase activity appears to be especiallyactive during the lipid modifications that arecoordinate with the formation of the membrane-bound electron transport system (28). Althoughthe proportions of eight fatty acids in phospha-tidylglycerol, phosphatidylethanolamine, and car-diolipin isolated from E. coli are similar, the levelof C17: C is significantly higher in the phosphatidyl-ethanolamine (5). The higher level of C17: Cappears significant as the level of C6: I is notdepressed. If the high C17:c represented a con-version of the C16: , the increase in C17:c wouldhave been accompanied by a decrease in C16: l.Significant differences in the fatty acid composi-tion between isolated phosphatidylethanolamineand cardiolipin have been detected in severalspecies of mycobacteria (14). In the photosyn-thetic bacterium Rhodopseudomonas capsulata,the fatty acid composition of the phosphatidyl-ethanolamine and phosphatidylcholine were simi-lar, but they differed remarkably from the fattyacid composition of the phosphatidylglycerol(31). In Chlorella vulgaris, isolated mono- anddigalactosyldiglyceride, sulfoquinovosyldiglyc-eride, phosphatidylglycerol, phosphatidylcholine,phosphatidylethanolamine, phosphatidylinositol,and cardiolipin differ significantly in the pro-portions of 10 fatty acids and in the changes in

proportions in these fatty acids between light- anddark- grown cells (12). Although the fatty acidcomposition of isolated lipids of only a few micro-bial species has been examined, the possibilitythat an active modification of the fatty acid com-position of the complex lipids after synthesisappears to be fairly common and suggests thattransferase activity might play an important rolein the matabolism of bacterial complex lipids.

ACKNOWLEDGMENTSThis study was supported by grant GM-10285 of

the Institute of General Medical Sciences, UnitedStates Public Health Service.

I wish to thank A. N. Tucker for her excellenthelp during this study, R. L. Lester and J. C. Dittmerfor their advice and encouragement, and J. H. Lawand R. L. Lester for the gifts of lipids.

LITERATURE CITED

1. Benson, A. A., and B. Maruo. 1958. Plant phos-pholipids. I. Identification of the phosphatidylglycerols. Biochim. Biophys. Acta 27:189-195.

2. Bligh, E. G., and W. J. Dyer. 1959. A rapidmethod of total lipid extraction and purifica-tion. Can. J. Biochem. Physiol. 37:911-917.

3. Folch, J., M. Lees, and G. H. Sloane-Stanley.1957. A simple method for the isolation andpurification of total lipids from animal tissues.J. Biol. Chem. 226:497-509.

4. Hagen, P. O., H. Goldfine, and P. J. L. B. Wil-liams. 1966. Phospholipids of bacteria with ex-tensive intracytoplasmic membrane. Science151:1543-1544.

5. Kanemasa, Y., Y. Akamatsu, and S. Nojima.1967. Composition and turnover of the phos-pholipids in Escherichia coli. Biochim. Biophys.Acta 144:382-390.

6. Kanfer, J., and E. P. Kennedy. 1964. Metabolismand function of bacterial lipids. II. Biosyn-thesis of phospholipids in Escherichia coli. J.Biol. Chem. 239:1720-1726.

7. Kates, M. 1956. Hydrolysis of glycerolphospha-tides by plastid phospholipase C. Canad. J.Biochem. Biophys. 34:967-980.

8. Kates, M. 1960. Chromatographic and radio-isotopic investigation of the lipid componentsof runner bean leaves. Biochim. Biophys.Acta 41:315-328.

9. Lane, E. S. 1965. Thin layer chromatography oflong-chain tertiary amines and related com-pounds. J. Chromatog. 18:426-430.

10. Lester, R. L., and D. C. White. 1967. Quantita-tive gas-liquid chromatographic analysis ofethanolamine, monomethyl ethanolamine, anddimethyl ethanolamine from lipids. J. LipidRes. 8:565-568.

11. Macfarlane, M. 1962. Characterization of lipo-amino-acids as o-amino-acid esters of phos-phatidyl-glycerol. Nature 196:136-138.

12. Nichols, B. W. 1965. Light induced changes inthe lipids of Chlorella vulgaris. Biochim. Bio-phys. Acta 106:274-279.

1169VOL. 96, 1968

on May 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACrERIOL.

13. Novak, M. 1965. Colorimetric ultramicro methodfor the determination of free fatty acids. J.Lipid Res. 6:431-433.

14. Okuyama, H., T. Kankura, and S. Nojima.1967. Positional distribution of fatty acids inphospholipids from mycobacteria. J. Biochem.(Tokyo) 61:732-737.

15. Smith, I. 1960. Chromatographic and electro-phoretic techniques. Vol. 1, 2nd ed., p. 245.Interscience Publishers, Inc., New York.

16. Waldi, D. 1965. Spray reagents for thin layerchromatography, p. 483-502. In E. Stahl (ed.),Thin layer chromatography. Academic Press,Inc., New York.

17. Wawszkiewicz, E. J. 1961. A two-dimensionalsystem of paper chromatography of some sugarphosphates. Anal. Chem. 33:252-254.

18. Wells, M. A., and J. C. Dittmer. 1963. The useof Sephadex for the removal of nonlipid con-taminants from lipid extracts. Biochemistry2:1259-1263.

19. Wells, M. A., and J. C. Dittmer. 1966. A micro-analytical technique for the quantitative de-termination of twenty-four classes of brainlipids. Biochemistry 5:3405-3418.

20. White, D. C. 1962. Cytochrome and catalasepatterns during growth of Haemophilus parain-fluenzae. J. Bacteriol. 83:851-859.

21. White, D. C. 1964. Differential synthesis of fiveprimary electron transport dehydrogenases inHaemophilus parainiluenzae. J. Biol. Chem.239:2055-2060.

22. White, D. C. 1965. Synthesis of 2-demethylvitamin K2 and the cytochrome system inHaemophilus. J. Bacteriol 89:299-305.

23. White, D. C. 1965. The function of 2-demethylvitamin K2 in the electron transport system inHaemophilus parainfluenzae. J. Biol. Chem.240:1387-1394.

24. White, D. C. 1966. The obligatory involvementof the electron transport system in the catabolic

metabolism of Haemophilus parainfluenzae.Antonie van Leeuwenhoek J. Microbiol. Serol.32:139-158.

25. White, D. C. 1967. Effect of glucose on the for-mation of the membrane-bound electron trans-port system in Haemophilus parainfluenzae.J. Bacteriol. 93:567-573.

26. White, D. C., and R. H. Cox. 1967. Identificationand localization of the fatty acids in Haemo-philus parainfluenzae. J. Bacteriol. 93:1079-1088.

27. White, D. C., and F. E. Frerman. 1967. Extrac-tion, characterization, and cellular localizationof the lipids of Staphylococcus aureus. J.Bacteriol. 94:1854-1867.

28. White, D. C., and F. E. Frerman. 1968. Fattyacid composition of the complex lipids ofStaphylococcus aureus during the formation ofthe membrane-bound electron transport sys-tem. J. Bacteriol. 95:2198-2209.

29. White, D. C., and L. Smith. 1964. Localizationof the enzymes that catalyze hydrogen andelectron transport in Haemophilus parain-fluenzae. and the nature of the respiratorychain system. J. Biol. Chem. 239:3956-3963.

30. Wichert, P. V. 1962. Enzymatische Bestimmungvon L-a-Glycerophosphat in normalen undbelasteten Warmbluterorganen. Biochem. Z.336:49-55.

31. Wood, J. B., B. W. Nichols, and A. T. James.1965. The lipids and fatty acid metabolism ofphotosynthetic bacteria. Biochim. Biophys.Acta 106:261-273.

32. Wright, E. A., and D. C. White. 1966. Formationof a functional electron transport system duringthe growth of pencillin-induced spheroplasts ofHaemophilus parainifluenzae. J. Bacteriol. 91:1356-1362.

33. Vorbeck, M. L., and G. V. Marinetti. 1965. Sepa-ration of glycosyl diglycerides from phospha-tides using silicic acid chromatography. J.Lipid Res. 6:3-6.

1170 WHITE

on May 7, 2021 by guest

http://jb.asm.org/

Dow

nloaded from