PROPERTIES OF STAPHYLOCOCCUSAUREUS · Molisch test was strongly positive, whereas the biuret test...

ANTIGENIC PROPERTIES OF STAPHYLOCOCCUS AUREUS

PER OEDINGUniversity of Bergen, School of Medicine, The Gade Institute, Department of

Microbiology, Bergen, Norway

I. Introduction ............................................................................II. Antigenic Structure......................................................................

A. What Can Be Learned from Attempts at Serologic Type Differentiation.................B. Carbohydrate A and Other Carbohydrate Substances..................................C. Verwey's Protein Antigen..............................................................D. Nucleoproteins, Heterophile Antigens.................................................E. Hemagglutination......................................................................

F. Agar Precipitation ...................................................................G. Distribution of Antigens and Chemical Structures in the Cell...........................H. Animal Strains .........................................................................I. "Normal" Antibodies ...................................................................

III. Serologic Typing..........................................................................A. Methods and Techniques .............................................................

1. Cowan's Method..............................................................2. Oeding's Method...............3. Technical Problems ........................................................

B. Epidemiologic Value..............................................................1. Reproducibility.....................................................................

2. Field Examinations ..................................................................3. Correlation to Disease...............................................................4. Correlation to Antibiotic Sensitivity................................................5. Correlation to Phage Types and Groups..............................................

IV. Conclusion ..............................................................................V. References ..............................................................................

I. INTRODUCTION

The antigenic properties of Staphylococcusaureus have been considered difficult to explore.Methods which were used with success on relatedbacteria, failed with staphylococci. There hasbeen a general feeling that the antigenic struc-ture of this organism is so complex that a sys-

tematic classification can hardly be achieved.Consequently most work has been done on thecultural and biochemical behavior of the or-

ganism, whereas there has been little interest inits serologic and immunologic properties.

Staphylococci are extremely widespread andS. aureus is one of the organisms most frequentlyfound in human infections. At present staphy-lococcal hospital infections are a world-wideproblem of major importance. Knowledge of theantigens of S. aureus and their activities is notmerely of interest for epidemiologic typing, butmay also give possibilities for successful researchin the field of immunology and infection.

Extensive reviews have recently been written

on S. aureus in general (20) and on its patho-genicity (5). As there is today an increasing in-terest in the antigenic properties of S. aureus, a

review may be of help to those working or plan-ning to work in this field. This review has beenlimited to the bacterial antigens of S. aureus andserologic typing, whereas toxins, pathogenicity,and immunity in man are not covered.

II. ANTIGENIC STRUCTURE

A. What Can Be Learned from Attempts atSerologic Type Differentiation

A serologic distinction between pathogenic andsaprophytic staphylococci was first made byKolle and Otto (53). They found that strainsfrom purulent lesions were all agglutinated in a

serum against one type, whereas strains fromother sources were not agglutinated. Severalauthors in the following years confirmed thatpathogenic and saprophytic staphylococci couldbe separated by agglutination. However, it soon

became apparent that some strains of pathogenic374

374374374376377378380381382384385385385386;386387388388389'390390391391392

on April 12, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

http://mmbr.asm.org/

ANTIGENIC PROPERTIES OF S. AUREUS

staphylococci agglutinated only weakly or not atall in sera against pathogenic staphylococci usedfor group-differentiation, whereas other strainsagglutinated weakly in heterologous groupserum also.Even from these early investigations important

conclusions can be drawn on the antigenic struc-ture of S. aureus:

1. S. aureus and S. albus1 strains have mainlydifferent antigens. At least one common agglutin-ogen seems to be present in some strains.This has later been confirmed by Oeding (66)and Pillet and Orta (88).

2. S. aureus strains have at least one commonagglutinogen. By agglutinin absorption it wasshown (Hine (35); Julianelle (44); and laterworkers) that S. aureus has a very complexantigenic structure with a number of more orless shared antigens.

In most earlier studies, immunization and ag-glutination were performed with boiled bacteria.This was done partly because heating resulted infewer cross-reactions, thus facilitating serologicclassification. Andersen (2) and Oeding (66)tried to get a more complete picture of theantigens of S. aureus by including heat-labileantigens. Oeding (66), by immunizing withformalin-killed vaccines and making cross-absorptions and agglutinations with live bac-teria, worked out an antigenic scheme for S.aureus. Ten antigens were recorded and givenletters from a to k. At least one heat-stableantigen (d) is shared by all S. aureus strains.The introduction of an antigenic scheme was

considered by Oeding to be a necessary develop-ment which might stimulate further research onstaphylococcal serology. Other workers havefollowed this line (8, 25, 57, 105, 110), althoughthe results are usually not directly comparablebecause different strains and designations ofantigens have been used. Grun (25), however,using Oeding's technique and strains, confirmedthe results obtained by the latter. Grun (28)

1 Editor's note: Since in many cases one can notascertain which of the present designations,Staphylococcus aureus or Staphylococcus epidermi-dis (Breed, R. S., Murray, E. G. D., and Smith,N. R. 1957 Bergey's manual of determinativebacteriology, 7th ed. The Williams & WilkinsCo., Baltimore), fits strains of staphylococcidesignated as Staphylococcus albus and Staphy-lococcus citreus in the literature, the older specificepithets are retained in this review.

also described a new antigen which he designatedas 1, present chiefly in animal strains but also inhuman ones. Haukenes and Oeding (33) recentlydiscovered by agglutinin absorptions and by agarprecipitation that the e antigen consists of twodistinct antigens and that there is also a previ-ously unrecognized antigen present in the strainsexamined. The new antigens were given theletters m and n.

Strains are characterized by their antigenicpatterns, and groups and types by the similarityof antigenic patterns and not by antigens char-acteristic of each group or type. Thus in staphy-lococci one can hardly speak of type-specificantigens. Major and minor antigens may bemore adequate terms.When suspensions of staphylococci are heated,

a significant reduction in the agglutinatingability of the antigens in factor sera is firstobserved at 100 C, and then a reactivation at120 C (68). Naturally this may be of great con-sequence in the determination of heat-stableantigens in staphylococci. The observation wasat first difficult to explain, but when the signifi-cance of blocking antigens became apparent, itseemed to be connected with this.

It was demonstrated (3, 68) that one antigenmight block agglutination of other antigenspresumably situated deeper in the cell. Inag-glutinability due to blocking antigens was shownto be of great consequence in the determinationof the antigens of a strain (Oeding (72)). Al-though any antigen, and possibly also inertmaterial, may block agglutination, the i antigenis usually the cause. As a rule the blocking effectis insignificant in cultures 3 to 5 hr old. It in-creases with the duration of incubation and maybe complete in cultures 18 to 24 hr old. Further-more, blocking increases when incubation takesplace at temperatures lower than 37 C and de-creases at higher temperatures. The discovery ofblocking antigens also gave an indication of thedistribution of the antigens in the cell (68).

It has been pointed out that extensive quanti-tative variation of the staphylococcal antigensfrom one strain to another makes classificationdifficult (36, 66). The ability of the antigens tomediate agglutination and to produce antibodiesalso varies from one strain to another. In onestrain an antigen may be a good agglutinogenbut poorly antigenic (66, 105). We have alsoobserved that although a weak antigen may not

1960] 375


http://mm

br.asm.org/

Dow

nloaded from


PER OEDING

be recognized by agglutination, it can absorbthe homologous antibody completely.

According to the sensitivity to heating and tothe action of proteolytic enzymes (3, 68, 89), animpression of the nature of the antigens was ob-tained. The resistant antigens were supposed tohave a carbohydrate structure, whereas thelabile ones would be proteins or possess deter-minate groups of protein nature. The majorityof the antigens recorded by Oeding (68) werefound to be heat-stable and possibly carbo-hydrates, for example the d antigen shared byall S. aureus strains. Pillet et al. (89) consideredthe agglutinogens to be chiefly of protein nature,a view which is now shared in this laboratory inthe light of recent observations. Oeding's eantigen was definitely heat-labile and was sup-posed to be a protein. This is also the case withthe newly described antigens m and n (33).However, certain antigens (a, b, c) seem to haveone heat-labile and another heat-stable com-ponent (Oeding (68)). This was also demon-strated by GrIn (29) in his new I antigen. Recentinvestigations indicate that this peculiar ob-servation may be explained by the presence inthe factor sera of more than one antibody (33).

B. Carbohydrate A and Other CarbohydrateSubstances

Julianelle and Wieghard (46, 47) demonstratedthat S. aureus and S. albus each had a carbohy-drate antigen shared by all strains in the twogroups, respectively. The precipitinogen presentin S. aureus was named carbohydrate A, thatpresent in S. albus, carbohydrate B. There wasno precipitation in sera against pneumococci,Klebsiella, or typhoid bacilli.

Staphylococci were extracted with hot 0.0625 Nhydrochloric acid, and all material precipitablewith 40 per cent NaOH and subsequently with 50per cent trichloroacetic acid was removed. TheMolisch test was strongly positive, whereas thebiuret test was variable. The total nitrogen wasabout 4 per cent. The high percentage of phos-phorus (about 6 per cent) suggested that thesubstances might be phosphoric acid derivatives.The two substances differed clearly in theiroptical rotation and in the sugars after hydroly-sis. The reducing sugars were not identified (113).The purified substances did not engender

antibodies when injected into rabbits, but pre-cipitated in dilutions up to 1 to 8 million in

homologous immune serum. The carbohydrateshad little ability to form antibodies even whenwhole organisms were injected. Cutaneous reac-tions of the immediate "wheal and erythematype" were observed with high dilutions ofcarbohydrate A, but not with carbohydrate Bin patients with staphylococcal infections.Circulating precipitins were present only in pa-tients with chronic and severe staphylococcalinfections (45, 46, 48).The existence of the two group-specific carbo-

hydrate precipitinogens of Julianelle and Wieg-hard has been confirmed by a number of authors(12, 107, 109). It was, however, observed thatthese substances were not present in all strainsof each group. S. albus strains lacking carbohydrateB were separated by Thompson and Khorazo(107) into one clearly defined and two less clearlydefined groups, and S. aureus strains lackingcarbohydrate A were placed in another group byCowan (12). In both instances very crude ex-tracts were used and it is not known whethersubstances specific for the new groups exist.The technique used by Julianelle and Wieghard

for extraction and purification is rather drastic.This fact, and the not insignificant nitrogencontent indicate that carbohydrate A was notproduced in its pure form. Serologically activecarbohydrates have been produced with signifi-cantly lower nitrogen values (73) or practicallynitrogen free (38, 108), but it is not knownwhether these substances contain the sameactive principle.

Various methods for extraction and fractiona-tion of carbohydrate materials from S. aureushave been tried. The procedure described byVerwey (109) seems to be comparatively gentleand was found to give satisfactory separation ofcarbohydrate, protein, and nucleoprotein. Theextract obtained after grinding the staphylococciis precipitated with 0.1 N hydrochloric acid, firstat pH 5.2 and then at pH 3.5. These two fractionscontain the bulk of the nucleoprotein. After aprotein fraction has been removed by precipita-tion with 50 per cent trichloroacetic acid, acarbohydrate is precipitated from the super-natant with alcohol. According to Verwey thissubstance, like carbohydrate A, had a faintlypositive biuret reaction, contained nitrogen andphosphorus but not pentoses, and was groupspecific. Similar substances were isolated byInoue (40) and Kahn et al. (49). The carbohy-

376 IVOL. 24


http://mm

br.asm.org/

Dow

nloaded from



drate isolated by Inoue was negative on thebiuret test, contained 3.5 per cent nitrogen and1.6 per cent phosphorus, and had a reducingpower after hydrolysis of 28.5 per cent. Pentoseswere present, whereas glucosamine was notdemonstrated. Kahn et al. considered the carbo-hydrate to be the most significant staphylococcalsubstance for cutaneous reactions in hypersensi-tive individuals.

Fellowes and Routh (22) and Oeding (71),examining various extraction methods, got themost satisfactory results with the hot formamidemethod. The carbohydrate obtained by Fellowesand Routh in the supernatant, after precipitationwith 2.5 volumes of alcohol, was obviously ratherimpure. It had a positive biuret reaction, con-

tained 6.5 to 8 per cent nitrogen, and inducedantibody formation when injected into rabbits.The amount of reducing sugars after hydrolysiswas about 20 per cent and pentoses and glucosa-mine were demonstrated. Oeding (71) isolatedtwo carbohydrates by alcohol fractionation ofthe formamide extract. The fraction which pre-cipitated with2 volumes of alcohol was apparentlythrown away by Fellowes and Routh. It pre-cipitated together with the protein when acidwas added and was nondialyzable, whereas thecarbohydrate precipitated by 4 volumes ofalcohol was partly dialyzable. These substancesdid not contain free protein but there was stillsome nitrogen and both contained pentoses. The4-volume substance may be identical withcarbohydrate A.

Sasaki (personal communication) also obtainedtwo carbohydrate substances from a formamideextract of one S. aureus strain. In both substancesglucose, galactose, xylose, and glucosamine were

identified by paper chromatography.Hoffstadt and Clark (38) and Maggi (60)

isolated different carbohydrates from smoothand rough variants of S. aureus. The rough-variant carbohydrate of Hoffstadt and Clark was

apparently antigenically inert.Jensen (42, 43) extracted from S. aureus by

means of phosphate buffer or heating to 100 C a

substance which was present in the majority ofS. aureus but not in S. albus strains and which was

toxic to guinea pig ileum. The substance was

labile on autoclaving and storage. The biurettest was negative and the substance was con-

sidered to be a carbohydrate. Few chemical data

are given. It seems that this product must be amixture (see below).

In this laboratory, staphylococci are nowfrozen, crushed in a bacteria press, and extractedwith phosphate buffer. A crude fraction con-taining the bulk of the carbohydrate is obtainedby acid or alcohol precipitation, and the materialis purified by ion exchange chromatography.The group-specific substance thus obtained isprobably identical with the active material incarbohydrate A. It is active in very high dilutionin the ring test, where a sharp disc is formed, anddoes not induce antibodies in rabbits or sensitizeerythrocytes to agglutination. A characteristicband is formed by agar precipitation. The biuretreaction is negative, the Molisch test very weaklypositive. The substance has a reducing activityafter hydrolysis of 30 per cent. Glucosamine andprobably related structures constitute an im-portant part of the substance, whereas othersugars have not been demonstrated by two-dimensional chromatography (Haukenes, un-published data).The investigations referred to above -how

that there are more antigens of carbohydratenature than the group A substance present instrains of S. aureus. Comparison between thecarbohydrates isolated by different authors isextremely difficult. Precipitation will differ ac-cording to the methods used for extraction andpurification, and with the concentration andpurity of the substance. Bindings may be re-leased with one method whereas new bindingsmay be formed with another method. Further-more, serologic investigations have shown thatthe numbers and quantities of carbohydrateantigens other than the d antigen vary fromstrain to strain. It is likely that the substancesdescribed have been mixtures of serologicallydifferent carbohydrates, possibly containingother material also. The true nature of carbo-hydrate A is still unknown.

C. Verwey's Protein Antigen

Verwey (109) demonstrated a small amount ofa protein fraction in S. aureus which -was groupspecific and serologically active in high dilution.After nucleoproteins had been removed byprecipitation with 0.1 N hydrochloric acid atpH 5.2 and pH 3.5, this protein was broughtdown by addition of 50 per cent trichloroaceticacid to a final concentration of 11 to 14 per cent.

37719601


http://mm

br.asm.org/

Dow

nloaded from


TABLE 1Some properties of carbohydrates and proteins derived from Staphylococcus aureus

Reference Molisch Biuret N P Reducing Glucos- Gel-Diffusion NatureTestTest Sugar amine

Julianelle and Strong Variable 4.1 6.3 26 CarbohydrateWieghard (46, (A), group47) specific

Hoffstadt and Strong Negative 0.1-0.3 0 CarbohydrateClark (38)

Inoue (40) Strong Negative 3.5 1.6 28.5 0 CarbohydrateVerwey (109) Strong Weak Present Present Carbohydrate,

group specificFellowes and Positive Weak 6.5-8.2 1.0-1.2 18.3-21.6 6.7-9.9 CarbohydrateRouth (22)

Haukenes (un- Weak Negative 2.5 6 30 19 Charac- Carbohydrate,published data) teristic group specific

band

Verwey (109) Weak Strong 14 <0.1 Protein, groupspecific

Inoue (40) Positive Positive 11 1.5 8.4 Positive ProteinDworetzky et Weak Positive 12.1 1.3 Protein

al. (18)

The substance was strongly positive to thebiuret test and weakly positive to the Molischtest. It contained about 14 per cent nitrogenand less than 0.1 per cent phosphorus, and wasdestroyed by trypsin. Pentoses were not present.

Similar substances have been isolated byInoue (40) and Kahn et al. (49), although theygave slightly lower values for protein and highervalues for phosphorus. The material describedby Inoue contained glucosamine. Its serologicspecificity was not described. The protein ma-terial of Kahn et al. was reported to be serologi-cally highly specific and to give skin reactionsboth in normal and hypersensitive individuals.Sasaki (personal communication) examined severalprotein fractions from yellow and white staphy-lococci by paper chromatography and foundthem to be almost homogeneous.Very little chemical work has therefore been

done on the protein antigens of S. aureus. Theprotein of Verwey seems to be a clear-cut entitybut little is known of its chemical structure orimmunologic significance. Although it is said tobe group specific, we do not know to what extentit is present in strains of S. aureus. By serologicinvestigations several antigens of protein naturehave been found in S. aureuis. These proteins

are, however, not group antigens but more orless "specific." Verwey's protein may be a mix-ture from a serologic point of view, or otherprotein antigens may have been lost, probablyinto the bulk of nucleoproteins. Furthermore,the possibility of isolating protein antigens willdepend upon the strains used.

D. Nucleoproteins, Heterophile AntigensSerologically active nucleoprotein materials

have been isolated from staphylococci by anumber of authors. The nucleoproteins wereextracted with 0.01 N NaOH and precipitatedwith dilute acetic acid by Lancefield (55), Boorand Miller (6), and Julianelle and Wieghard(48). Verwev (109) obtained the bulk of thenucleoproteins by precipitation with 0.1 N hydro-chloric acid at pH 5.2 and then at pH 3.5. Thisprocedure wAas adopted by Kahn et al. (49) andDworetzky et al. (18). The latter authors demon-strated 12.5 per cent nitrogen and 3.7 per centphosphorus in the nucleoprotein fraction of onestrain of S. aureus, and they as well as Boor andMiller found indications of a carbohydrateradical.

Lancefield (55) reported cross-precipitationbetween nucleoproteins from staphylococci,

378 PER OEDING tVOL. 24


http://mm

br.asm.org/

Dow

nloaded from



TABLE 2Cross-reactivity of "nucleoproteins" derived

from Staphylococcus aureus

1. Extraction with 0.01 N NaOH, precipitationwith acetic acid.

2. Species-specific skin reactivity.3. Cross-reactions with streptococci; pneumo-

cocci.

(Lancefield (55) and others.)

1. Extraction after grinding, precipitation withacetic acid pH 5.2 and pH 3.5.

2. Cross-reaction with Staphylococcus albus.(Verwey (109) and others.)

streptococci, and pneumococci, but not fromEscherichia coli or Corynebacterium diphtheriae.Similar observations were made by Boor andMiller (6), who found precipitation of S. aureus

nucleoprotein in sera against pneumococci butnot in sera against Neisseria species. It has beenstated that pathogenic and saprophytic staphy-lococci have a common nucleoprotein antigen(17, 48, 109), and that the nucleoproteins pre-

cipitated at pH 5.2 and pH 3.5 cross-react (49).Furthermore, these nucleoproteins showed a

broad species-specific skin reactivity in contrastto the A carbohydrate, and the reaction was ofthe delayed inflammatory type (48, 49). AlthoughJulianelle and Wieghard (48) observed no toxiceffect in rabbits or white mice, Dworetzky et al.(16) found the nucleoprotein toxic to the ileumof guinea pigs.

Savino (98) by precipitation with alcohol andacetic acid, isolated a substance which was

supposed to be a carbohydrate but which, ac-

cording to the chemical data given, is more

likely to have contained nucleoproteins. Thisstaphylococcal substance precipitated in antiseraagainst staphylococci, pneumococci, anthraxbacilli, and meningococci.However, with the technique used by Lance-

field (55) and others, the nucleoproteins are

probably split at an alkaline reaction. Further-more, by the addition of acid, various proteinantigens will precipitate out at their isoelectricpoints. The described nucleoproteins are there-fore very likely to be mixtures of antigens ofdifferent chemical nature and serologic specificity.The cross-reactivity may be due to a substanceother than a nucleoprotein.

Staphylococci have been found to absorbantibodies to blood groups A and B from humansera (39). This reaction was thought to be dueto polysaccharides. The sharing of a commonantigen in these species is very unlikely, sincethey do not cross-agglutinate (20).

Cross-reactivity between staphylococci andListeria monocytogenes has been demonstratedby Seeliger and Sulzbacher (100). Of the staphy-lococcal strains examined, 30 to 50 per cent gaveagglutination and complement fixation in liveand in autoclaved conditions with antiseraagainst live L. monocytogenes types 1, 2, and 3.Soluble carbohydrate materials gave only aunilateral cross-reaction between L. mono-cytogenes type 3 serum and staphylococcal ex-tracts. The latter result was confirmed using anagar precipitation technique. Three separatebands were found to be common to the abovementioned system. Parts of the antigenic com-plexes other than the extractable carbohydratematerial must therefore be responsible for theextended cross-reactions observed with wholebacteria. The nature of the cross-reacting sub-stance is unknown (100).

In the course of purification of staphylococcalextracts by column chromatography, Haukenes(unpublished data) isolated one contaminatingstrain of Alcaligenes and one of Bacillus subtilis,both of which gave bands with staphylococcalrabbit antiserum by agar precipitation.

Mitchell and Moyle (62, 63) discovered thatlipid-free extracts of staphylococci containedphosphorus, amounting to more than 30 per centof that which could be accounted for by thenucleic acid content. The excess phosphate wasseparated from the other phosphorus compoundsafter fractional extraction of staphylococci indilute acid or alkali. Excess phosphate wasdemonstrated in a number of gram-positiveorganisms, suggesting that a glycerophosphatecompound similar to that of S. aureus may be ofgeneral occurrence in gram-positive bacteriaand yeasts.

Recently McCarty (58) recognized in strepto-cocci a cross-precipitating substance whichcould be extracted at pH 2 at 100 C, at pH 9, bytrypsin treatment, or with hot formamide. Thesodium salt of the purified antigen was essentiallynitrogen free and on the border line of dialyza-bility through cellophane. Paper chromatographyof acid hydrolyzates showed no carbohydrate or

1960] 379


http://mm

br.asm.org/

Dow

nloaded from


PER OEDING

TABLE 3Cross-reactivity observed between Staphylococcus

aureus and:

1. Human blood groups A and B.(Illchmann-Christ and Nagel (39).)

2. Listeria monocytogenes.(Seeliger and Sulzbacher (100).)

3. Alcaligenes, Bacillus subtilis.(Haukenes, unpublished data.)

4. Staphylococcus albus, streptococei, pneumo-

cocci, B. subtilis. Hemagglutination, non-

species-specific substance.(Rantz et al. (92, 93).)

5. Glycerophosphate extracted from streptococci.Reacts also with S. albus and Bacillus species.(McCarty (58).)

amino acids. A single organic phosphate was

demonstrated, its mobility being identical withthat of glycerophosphate. Synthetic polyglycero-phosphates inhibited the precipitin reaction.The substance was estimated to constitute about1 per cent of the dry weight of the streptococcalcells.

This antigen was found in many differentgroups of streptococci, including nonhemolyticspecies, in S. aureus and S. albus, and in aerobicsporulating bacilli including B. subtilis. It was

not present in pneumococci, clostridia, coryne-

bacteria, yeasts, ilicrococcus lysodeikticus, or

Staphylococcus citreus.There is not complete accordance between the

distribution of the excess phosphate of Mitchelland Moyle and the phosphate antigen ofMcCarty. The explanation is probably that theantigenic substance constitutes only a part ofthe excess phosphate. The phosphate antigendoes not precipitate with acid and can thereforehardly be present in the nucleoproteins ofLancefield and others.Three observations have been described which

are difficult to explain but may be due to hetero-phile antigens. Dudgeon and Bamforth (15)found precipitins in no less than 87 per cent ofthe sera from patients with serum sickness againsta 1-month-old broth culture filtrate of S. aureus.

No precipitation was seen with extracts of E. colior Salmonella typhosa. There may be a connectionwith the serologic similarity demonstrated be-tween staphylococci and L. nionocytogenes (100),glandular fever (23), and blood groups (39).

Penner and Voldrich (81) found that S. aureus

agglutinated in high titers in all tuberculouspleural effusions examined, and less frequentlyin nontuberculous effusions. There was no ag-glutination of S. albus or S. citreus, or of otherbacteria. The agglutinating substance was notdestroyed by heat, whereas the factor present inthe effusions was inactivated at 60 C. Shrigley(101) observed that approximately 46 per centof S. aureus strains and possibly strains of S.albus, but not of S. citreus or other organismstested, were agglutinated in normal or influenzavirus-infected chorioallantoic fluid. The fluidswere rendered inactive after heating to 56 C. Theactive substance was nondialyzable and con-sidered to be a protein.

There is a possibility that these two observa-tions can be explained by the same mechanism.The tuberculous effusions and the chorioallantoicfluid are similar in so far as both agglutinate S.aureus exclusively and both are highly susceptibleto heat. Although the agglutination may be hardto explain in terms of antigen-antibody reactions,some kind of heterophile antigen may be in-volved. Heterophile antigens in S. aureus,demonstrated by indirect hemagglutination, willbe treated in the following section, after whicha concluding remark will follow.

E. HemagglutinationKeogh et al. (50) found that staphylococci

among other bacteria were able to sensitizeerythrocytes to agglutination in the presence ofantiserum. Later authors confirmed that culturefiltrates (64, 93) and extracts of S. aureus haveerythrocyte sensitizing properties. Active ma-terial was present in the insoluble fraction afterextraction with 90 per cent phenol (50, 73, 94),and in the soluble fraction after extraction with88 per cent phenol (93). Furthermore activematerial could be extracted with glycine (34),hot formamide (73, 95), hot hydrochloric acid(93), by simple heating to 100 C, or by means ofproteolytic enzymes (78).

Little has been done to isolate the activeprinciple. Most investigations on the propertiesof the sensitizing substance have been madewith simple culture filtrates or very crude ex-tracts. It is generally assumed that the sensitizingsubstance of staphylococci, as of other bacteria,is a polysaccharide, although Sachse (95) con-siders it to be a nucleoprotein. The sensitizingproperty has been shown to resist autoclaving

380 IVOL. 24


http://mm

br.asm.org/

Dow

nloaded from



(59, 73) and formamide extraction at 150 C(73, 95). On the other hand, the sensitizing sub-stance with heterophile properties described byRantz et al. (92) was shown to be stable at 100 Cat acid pH, but labile at neutrality and alkalinepH. This was confirmed by Neter and Gorzynski(64). At room temperature the sensitizing prop-erty was destroyed both by acid and alkali (73).The sensitizing substance is not destroyed bytrypsin (59, 73, 92) and is nondialyzable (78, 92).

Oeding (73), working with more purified ma-terials, found them almost nitrogen-free. Thesensitizing property was gradually lost onstorage in an ice chest (73, 94). According toOeding (73) this indicates a labile grouping inthe active polysaccharide, necessary for itssensitizing activity but without significance toprecipitation, which was unimpaired.

Antibodies to the sensitizing antigen wereeasily produced in rabbits when crude materialwas injected. Purified material, which is notantigenic, becomes so when attached to eryth-rocytes (73).The reports of various authors are quite con-

flicting regarding the properties of the sensitizingsubstance. This is not surprising as differentstaphylococcal strains and even different bac-terial species have been compared. Furthermore,the sensitizing substance may have differentproperties in crude preparations than in purifiedextracts. The investigations hitherto reportedcan hardly tell us whether more than one sensi-tizing antigen is present in S. aureus and it canbe discussed whether it is justified to describethe properties, as has been done here, as if wewere dealing with only one substance.However, the sensitizing agents described by

the majority of authors show obvious similaritiesin that they are apparently common to staphy-lococci and certain other bacterial species. Asensitizing substance common to S. aureus andS. albus strains has been demonstrated (59, 64,94), which is apparently present also in strepto-cocci, pneumococci, and B. subtilis, but not inC. diphtheriae, S. citreus, M. lysodeikticus, orgram-negative bacteria (34, 64, 78, 93, 95), withthe exception of one Pseudomonas strain (93).Antibodies to the sensitizing substance weredemonstrated in the sera of three patients withglandular fever (23), which is interesting in rela-tion to the antigenic similarity demonstratedbetween staphylococci and L. monocytogenes

((100) see above). Rantz et at. (92, 93) havepayed particular attention to the heterophilesensitizing antigen in their extensive study, andnamed it NSS, i.e., nonspecies-specific substance.

Neter and Gorzynski (64) obtained highertiters by indirect staphylococcal hemagglutina-tion when the erythrocytes were treated withenzymes. The difference in titer between enzyme-treated and nontreated erythrocytes was thoughtto be due to incomplete antibodies. Sensitizingmaterial was present in varying quantities indifferent strains of staphylococci.

In this connection it should be mentioned thatKourilsky et al. (54), by means of the immune-adherence test, demonstrated an antigen-anti-body system which was different from theagglutinogen-agglutinin system. The specificreceptor was present only in human erythrocytesand the active substance was considered to be alipopolysaccharide or a polysaccharide.When the properties of the different substances

which have been described as giving cross-reac-tions, either by precipitation or by hemagglutina-tion, are compared, it may seem improbable thatonly one heterophile antigen should be present inS. aureus. It is striking, however, that the speciesreactivity of the "nucleoprotein" of Lancefieldand others is very similar to that of the eryth-rocyte-coating substance described above. Cer-tainly related species such as staphylococci,streptococci, and pneumococci may have morethan one antigen in common, and it cannot bedenied that the "nucleoprotein" and the sensi-tizing substance show significant differences.But it should be remembered that the conditionsfor characterization of the substance have beenrather unsatisfactory.

This is a very interesting and important fieldfor further study. It will be necessary to makecomparative studies of heterophile substances bydifferent antigen-antibody reactions, and specialattention should be given to the comparison ofthe heterophile "nucleoprotein," the heterophilephosphate of McCarty, and the heterophilesensitizing substance.

F. Agar PrecipitationThe double diffusion technique in agar pro-

vides a very refined way of separating differentantigen-antibody systems. Investigations haveshown that in culture filtrates or extracts of S.aureus, a large number of bands, each probably

38119601


http://mm

br.asm.org/

Dow

nloaded from


PER OEDING

representing a distinct antigen-antibody system,can be demonstrated. Some of the bands aredue to toxin-antitoxin systems, others apparentlyto bacterial antigen-antibody systems (4, 42, 74,80, 105). The number of bands depends uponwhether bacterial filtrates or extracts are used(4, 11). A correlation has further been demon-strated between the number of lines and thevirulence of the staphylococcal strains.

Certainly the mere demonstration of a numberof bands in a staphylococcal system has in itselfa very limited value unless the bands can beidentified. Stern and Elek (105) demonstrated asingle band in acid extracts of S. aureus cellwalls which was apparently due to a groupantigen, whereas the endoplasm showed a numberof bands which seemed to represent both group-and type-specific antigens. Jensen (42, 43)extracted from S. aureus, by means of phosphatebuffer or heating to 100 C, a substance whichformed a very strong band in agar. Precipitatingantibodies to this substance were found in veryhigh titers in all normal human sera examined.Personal investigations ((74) and unpublisheddata) have shown that a number of bands, someof which are closely situated, are common tostrains of S. aureus. The identity of these groupbands is unknown. The description and pictures,presented by Jensen, give the impression thatthe group band, which is very wide, must consistof at least two or three separate systems.The antigens forming the group bands of S.

aureus do not seem to be present in S. albus,although we have occasionally observed weakbands common to the two groups. The observa-tion by Haukenes (unpublished data) of a bandcommon to S. aureus, B. subtilis, and an Al-caligenes sp., also indicates that the heterophileantigens present in S. aureus can be studied bymeans of the agar precipitation technique.

It was shown bv this technique that certainantigens are removed by washing the staphy-lococci (Haukenes, unpublished data) and thatunwashed bacteria, disrupted in a frozen condi-tion in a bacterial press, form the best materialif one is to obtain a complete antigenic picture.By comparing the bands with agglutination

systems using whole sera or the described factorsera, we have been able to identify some of theagglutinogens as distinct bands. Furthermore,staphylococcal extracts have been comparedby the ring test and by agar precipitation, and

the latter method has proved very useful infollowing the purification of extracts.However, for various reasons, it seems im-

possible to demonstrate all agglutinogen-ag-glutinin systems by agar precipitation, or toidentify all precipitation bands by agglutination.Thus the number of staphylococcal antigens isstill greater than shown by either of the tw-omethods. The separation of the e antigen intotwo distinct antigens (see above) illustrates thisfact. It was by means of agar precipitation thatthe existence of tw-o antigens was first suspected.One of the antigens (n) formed a characteristic,sharp line in agar, whereas the other antigen(el) formed no line.We feel that agar precipitation is a very

convenient and valuable method in staphylococ-cal research. However, the method undoubtedlyhas its limitations. One difficulty is the interpre-tation of the different bands, because it seemsthat the specificity of the method is so high thatantigen-antibody systems with only minimaldifferences form their ow-n bands. Small varia-tions in an antigen or antibody, which are prob-ably insignificant when the conventional antigen-antibody reactions are used, may easily takeplace in response to chemical treatment or evenstorage. The frequency with which "twin bands"can be produced is in our experience an indica-tion of the "over-sensitivity" of the method.

It is very important that the conditions forreaction of a bacterial extract and an antiserumbe thoroughly varied; for example, the concen-trations of the two components and the arrange-ment of the adjacent systems. Absorption testsshould also be included. It is our experiencethat unless these conditions are fulfilled, anantigen-antibody system may easily be over-looked.

G. Distribution of Antigens and ChemicalStructures in the Cell

The demonstration of a number of differentagglutinogens in intact cells of S. aureus suggeststhat these antigens are situated on the cellsurface or penetrate to it from deeper layers.Thus the outer surface of S. aureus must be verycomplex in its antigenic equipment. Blockingantigens may develop which render the cellsmore or less inagglutinable by specific antisera(3, 68). The development of blocking antigensand inagglutinability is of regular occurrence,

382 [VOL. 24


http://mm

br.asm.org/

Dow

nloaded from



depending upon the antigenic pattern of a strain(Oeding (72)). According to the general view,the blocking antigen probably forms a surfacelayer which covers the previously exposedantigens. Repeated washings are of little conse-quence (72). The surface layer of blockingantigen seems therefore to be firmly bound tothe cell and may be formed by the selectivedevelopment of one particular antigen in com-parison to the other antigens present on thesurface. The blocking antigen does not seem tobe destroyed by autoclaving, which neverthelessmakes the strains more agglutinable. The forma-tion of the outer blocking layer is dependent onincubation time, temperature, and probablyother factors.

Inagglutinability of staphylococci seems to bemore complex and in many ways different fromthe 0 inagglutinability caused by the Vi antigen.In staphylococci it is possible that every antigencan act as a blocking substance, provided it ispresent on the surface in sufficient quantity.Usually, however, the i antigen acts as the block-ing substance. Brodie (7) observed that the re-moval of the antibodies to the heat-stableantigens somewhat enhanced agglutination ofthe heat-labile antigens. Oeding (72) discussedthe possibility that a substance inhibiting ag-glutination is transmitted from the bacteria tothe serum during absorption.

Oeding (68) suggested the way the antigenshe described might be arranged, taking agglutina-tion of live and autoclaved bacteria and blockinginto consideration. There seems to be no generallaw as to the arrangement of the different antigensin the cells, although there is undoubtedly asimilarity between strains of the same serologicgroup. One antigen present on the surface andacting as a blocking substance in one strain, maybe present deeper in the cell and be blocked byanother antigen in a second strain. In somestrains the cells seem to be built up of the sameantigen, both in the deeper layers and on thesurface, as all antigens are active under differentexperimental conditions. This seemed to be thecase in the strains examined by Stern and Elek(105), who apparently found no antigenic dif-ference on agglutination between the whole celland the cell wall or between the inner and outersurface of the latter. Other strains seem to havelayers of different antigens; for example, a surfacelayer of a blocking antigen, then a layer of two

or three other antigens, and occasionally an innerlayer of an antigen which can be traced only incertain circumstances which are difficult topredict.Cohen et al. (11) demonstrated more antibody

components in sera against cellular protoplasmthan in sera against whole cells. This is difficultto explain unless the ultrasonic treatment usedto prepare protoplasm resulted in degradationand formation of new antigenic structures.

Apart from the very rare mucoid variants,staphylococci are not supposed to have truecapsules. Lyons (56) claimed that very youngcells of S. aureus have capsules containing atype-specific antigen, but this could not beconfirmed (52, 103). More recently Price andKneeland (90) demonstrated capsular swellingin young cells of a strongly mucoid strain of S.aureus 1 to 3 hr after the addition of homologousantiserum. Capsular swelling was also observedin the majority of nonmucoid strains of S. aureus,but not in S. albus strains, when antimucoid serumwas added (91). After prolonged immunizationeven nonmucoid strains produced antibodieswhich gave swelling with the homologous bac-teria. The capsular swelling antibody was ap-parently common to mucoid and nonmucoidstrains.

Although we have not examined the antibodyresponse to very young cultures of S. aureus,the results of Price and Kneeland are difficult toexplain in light of our observations that in youngbacteria a number of antigens are exposed on thesurface. A surface layer of blocking antigen hasnot yet developed. We have, however, recentlyobserved a substance in 18-hr cultures, by ringtest and agar precipitation, which seems to be aheterophile polysaccharide antigen. It is readilyremoved by washing and seems to constitute aloose surface layer (Haukenes, unpublished data).Although the capsule antigen of Price and Knee-land is described as group specific, there may bea connection between the two observations.Capsular swelling in nonmucoid strains wasdistinct only after 18 hr contact with antiserum.This might also indicate that the capsularswelling antigen is not exposed on the surfaceof the live cell, but is brought there by autolysis.

During recent years some interesting investi-gations have been made on the chemical structureof the cell wall of S. aureus. According to Stacey(104), stripping with sodium cholate indicates

3831960]


http://mm

br.asm.org/

Dow

nloaded from


PER OEDING

that the surface material of gram-positive bac-teria consists mainly of polysaccharides inpartial combination with nucleic acids, mainlyof the ribonucleic acid type, and relatively smallamounts of protein or fats. It has been estab-lished that the cell walls of many gram-positivebacteria possess distinctive properties, that anumber of the uncommon constituents of thecells are localized there, that cell wall peptidesare composed of a small variety of amino acids,and that amino sugars form important con-stituents of the cell wall (Salton (97)). The cellwalls of gram-positive bacteria are less complexin their amino acid and amino sugar composi-tion than those of gram-negative bacteria.

In S. aureus, small amounts of fats are presentin the cell wall. Dyar (19) estimated the amountof lipid to be about 10 per cent, present asphospholipid firmly bound to the cell surface.From electrophoretic studies of S. aureus beforeand after treatment with ribonuclease, it appearsthat nucleic acids are not located on the externalcell surface (32).Park and Strominger (79), having demon-

strated that three uridine nucleotides accumu-lated in a strain of S. aureus that was inhibitedby penicillin, found a structure analogous topart of the nucleotide in the cell wall. Thiscompound was supposed to be a precursor of thecell wall. A hexosamine identical to that de-scribed by Strange and Powell (106) was foundin the cell wall and in the nucleotide. This hasbeen shown to be a characteristic component ofcell walls of gram-positive bacteria by Cumminsand Harris (14). These authors also demon-strated the presence of glucosamine in the cellwall of S. aureus, whereas hexoses or pentoseswere not detected.The presence of a carbohydrate on the cell

surface of S. aureus was demonstrated by Dyar(19) and Webb (112). Serologic studies show thatantigenically active carbohydrates are presenton the surface of S. aureus (68).Webb (112) and Salton (96) conclude from

the results of enzymatic lysis of S. aureus thatprotein material is not normally exposed on thecell surface. Unheated bacteria are not lysed bytrypsin whereas heated bacteria are lysed, espe-cially when first treated with lysozyme. Thisconclusion is, however, based on the examinationof only two staphylococcal strains and maypossibly be valid for some strains and not forothers. Agglutination of live staphylococci has

shown antigens of protein nature to be presenton the cell surface, a typical example being thee antigen of Oeding (68). There is, therefore,hardly any doubt that protein antigens are fre-quently or regularly exposed at the cell surface.Mitchell and Moyle (63) are of the opinion thatthere is a preponderance of protein in the en-velope of S. aureus.A number of amino acids have been demon-

strated in whole cells of S. aureus. Alanine,glutamic acid, and lysine are present in highconcentrations in the cell walls of staphylococciand other gram-positive bacteria (14, 79). Inaddition, glycine seems to be regularly present inthe cell wall of S. aureus, and possibly certainother amino acids in lower concentrations (14,41, 63). Serine has been found, apparently inhigher concentration in S. albus than in S. aureusstrains (14). Park and Strominger (79) foundthat 45 per cent of the alanine and 92 per centof the glutamic acid present in the cell wall of S.aureus was in the D form, and similar findingsare reported by Salton (97).

Mitchell and Moyle (62) found that in thecell wall of S. aureus there is an excess of organicphosphate, the bulk of which is present in aneasily hydrolyzed glycerophosphate compound.Three-quarters of the weight of the cell wallfraction was estimated to be a glycerophosphate-protein complex (63). The presence of "excessphosphate" in all the gram-positive organismsexamined, but not in the gram-negative ones,suggests that glycerophosphate compounds arecommon structures of the cell walls of all gram-positive organisms.McCarty (58) recently demonstrated a

polymer of glycerophosphate with serologic ac-tivity in various gram-positive bacteria. Thecharacteristics of the purified antigen have beendescribed previously. The antigenic substancewas, however, not present in the cell wall frac-tion, and it appears that it accounts for only apart of the "excess phosphate" demonstrated byMitchell and Moyle.

H. Animal StrainsThe majority of the studies on the antigenic

structure of S. aureus have been performed onstrains of human origin and little is known of theanimal strains. However, it seems to be estab-lished that animal coagulase-positive staphy-lococci have their own antigenic patterns andfew antigens in common with human strains

384 [VOL. 2A


http://mm

br.asm.org/

Dow

nloaded from



(28, 61, 83). In absorbed or unabsorbed seraproduced against human staphylococci, only asmall number of animal strains are agglutinated,and if so usually weakly. Cowan (12), usingprecipitinogens prepared according to Julianelleand Wieghard (47), found animal strains tobelong mainly to group A. This should indicatethat some of the animal staphylococci havepolysaccharide A in common with humanpathogenic staphylococci. However, as pointedout above, the precipitinogens used were prob-ably rather impure, and the cross-reactions maywell be due to other common antigens.Grin (28) found the antigens described by

Oeding (66) infrequently represented in animalstrains. He demonstrated a new antigen (1)present not only in animal staphylococci butalso in human strains. This antigen was heat-stable and might be blocked in 24-hr cultures.

Pattison and Matthews (80) attempted todifferentiate animal coagulase-positive staphy-lococci by different antigen-antibody reactions.The presence of multiple common lines on agarprecipitation and the difficulty of absorptionindicate extensive sharing of antigens in strainsof widely different origin, as has been found alsoin human staphylococci. The slight antigenicdifferences observed may be related to hostspecies.

I. "Normal" AntibodiesIt has been repeatedly observed that nonim-

munized rabbits may have circulating antibodiesthat react with pathogenic staphylococci. Thetiter of the normal antibodies and its conse-quence for immunization and agglutination havebeen estimated very differently. Most authorshave found only low titers of "normal" anti-bodies and used a serum dilution of 1: 5 or 1 :10after immunization. Mercier et al. (61) are ofthe opinion that "normal" antibodies may con-fuse agglutination and prefer a serum dilution of1:50.In this laboratory, "normal" antibodies, re-

acting with our type strains of S. aureus, arefound regularly in rabbit sera. The titers areusually low (1:5 to 1:20) but titers up to 1:50are not infrequent. In the routine agglutinationtest, absorbed rabbit antisera are used in 1:10dilution and no disturbing effect due to "normal"antibodies has ever been observed. Probablythese "normal" antibodies are directed mainlyagainst the group antigens and therefore elimi-

nated from the immune sera during absorption.Pillet and Orta (86) found "normal" antibodiesin their rabbits mainly against types 3, 4, and8, and to a lesser degree against type 2, and con-sidered them to be directed against a commongroup antigen present also in animal staphy-lococci.

"Normal" antibodies against S. aureus presentin rabbit sera are considered to be due to previousinfections with staphylococci. We do not knowwhether the rabbit staphylococci are of animalor human origin. This is significant in view ofthe difference in antigenic structure of the twogroups (see above). Most probably the extentof staphylococcal infections and the productionof antibodies will differ from one laboratory toanother, and that may also be the case with thekind of antibodies produced.

Investigations on the presence of staphylococ-cal antibodies in human sera have mainly beenconcerned with antitoxic antibodies and verylittle with antibacterial antibodies. Rountreeand Barbour (94) demonstrated antibodies inhuman sera to the erythrocyte-coating substancepresent in pathogenic staphylococci. The anti-body was not transmitted from mother to child.Jensen (42, 43) found antibodies in very hightiters against one particular staphylococcal sub-stance in all human sera examined, and in humanmilk. Human y-globulin has been found to givemultiple bands by agar precipitation againstextracts of all staphylococcal strains examined(4), which indicates the presence of small amountsof different group antibodies in human sera.Neter and Gorzynski (64) demonstrated "in-complete" antibodies in high titer and "com-plete" antibodies to staphylococci in low titer inhuman y-globulin by hemagglutination.

Rantz et al. (93) found hemagglutinationantibodies against the heterophile substance NSSin human sera, and Dudgeon and Bamforth(15) observed that staphylococci were ag-glutinated by sera from patients with serumsickness. Thus "normal" antibodies reactingwith staphylococci are not necessarily specific ordue to previous staphylococcal infections.

III. SEROLOGIC TYPING

A. Methods and TechniquesPathogenic and saprophytic staphylococci

were differentiated serologically by Kolle andOtto (53) and later workers by agglutination in

3851960]


http://mm

br.asm.org/

Dow

nloaded from


PER OEDING

,sera against representative strains of the twogroups. The method was improved by the intro-duction of polyvalent S. aureus sera (9, 67). Adifferentiation can also be achieved using theprecipitation technique of Julianelle and Wieg-hard (12, 47, 107). The coagulase test has,however, made serologic methods for the charac-terization of pathogenic strains superfluous.

Until 1939 only scattered attempts had beenmade to classify S. aureus by serologic methods.It was realized that although precipitation couldseparate saprophytic from pathogenic staphy-lococci, a type division of the latter was notpossible with this technique (12). The comple-ment fixation test was also shown to be unsuit-able, although Seedorf (99), using absorbed sera,was able to distinguish 13 types among strains ofS. aureus.

In nearly all experiments in which a type dif-ferentiation of S. aureus was obtained, agglutininabsorption was used. When agglutination wasdone in tubes, from 3 to 7 types or groups werereported; one or two of them contained nearlyall typable strains while a great number ofstrains remained untypable (35, 44, 114).

1. Cowan's method. Cowan (12, 13) introducedthe slide agglutination technique, which givesmore distinct reactions than tube agglutination(36, 65). Boiled bacteria were used for the pro-duction of antisera, for absorption, and foragglutination. By means of absorbed and un-absorbed sera, Cowan classified strains of S.aureus into three types and an "atypical" group.Although similar results had been obtained byearlier authors (see above), Cowan's method wasthe first which proved to be serviceable for theclassification of S. aureus.By means of this technique, the number of

types was later extended. In addition to Cowan's3 main types, Gillespie et al. (24) registered 6subtypes and Christie and Keogh (10), 6 types.Hobbs (36) added 4 more types to the 9 ofCowan, and Christie and Keogh, the total num-ber of types thus amounting to 13. However,Mercier et al. (61) and Andersen and Heilesen(3) found no evidence of more than 3 typesamong the heat-killed international type strains,and considered further division to depend ex-clusively on quantitative differences.

Because of considerable antigenic overlapping,the technique and the determination of typeswere in many respects difficult. In particular

absorption presented a serious problem. First,the absorbed sera tended to have low titers,and second, in spite of this, cross-reactions wereusually found, i.e., pure type sera had not beenobtained. Indeed, Pillet and Orta (87) claimedto have obtained pure sera for all of Cowan's,and Christie and Keogh's 9 types, but the pro-cedure was difficult and there was no proof thatthe sera were monovalent.Hobbs (36) considered the strong reactions

to be due to major, and the weak reactions tominor, agglutinogens. According to her, theremay either be a large number of types capable ofrecognition by specifically absorbed sera, oronly a small number of specific types aroundwhich many strains showing minor variationsfrom the main type are grouped.

Strains were classed by the pattern of theirreactions (10) into types and subtypes, e.g.,la (24), or types 1 to 13 (36).Cowan's method has been used by a number of

authors and, although it has not been widelyadopted in routine epidemiologic investigations,some reports have given encouraging results(1, 21, 24, 37).In an epidemic of pemphigus neonatorum,

Andersen (2) was unable to obtain a satisfyingclassification of the strains by means of Cowan'smethod. However, when rabbits were immu-nized with trypsin-digested vaccines of selectedstrains, and trypsin-digested bacteria were usedfor absorption and agglutination, three antigens(a, b, and c) were demonstrated. A specificpemphigus serum was produced by means ofwhich the epidemic could be followed.

2. Oeding's method. In an epidemic of mastitis,Geding (65), by means of agglutinin absorption,obtained a serum containing the antibody spe-cific for the epidemic strain. Formalin-killedbacteria were used for the production of rabbitimmune sera and for absorption, and agglutina-tion was performed on slides with live 18-hr-oldbacteria.

Selected strains were now studied by meansof cross-absorption, on the basis of which anantigenic scheme was established (66). Tenantigens were determined and given letters froma to k. One of them (d) was common to all yellowstrains, and 2 (f, g) gave very weak antibodies.The resulting 7 antigenic factors were consideredto be distinct and the corresponding, apparentlymonovalent, factor sera were used for the sero-

386 [VOL. 24


http://mm

br.asm.org/

Dow

nloaded from



logic typing of pathogenic staphylococci. Strainswere designated by the factor sera in which theyagglutinated, the antigenic patterns being, forexample, abc, or that of the mastitis strain, ae.When it was discovered that antigens were

regularly blocked in 18-hr-old live bacteria, thestrains were also examined after autoclavingand designated by their antigenic patterns bothin the live and autoclaved condition (69).

Agglutination of live bacteria of the 13 inter-national type strains in "group" sera (a, b, c, e)showed that Cowan's types 1, 11, and III hadthe antigenic patterns abe, ab, and abc, respec-tively. These patterns were found to be widelydistributed and, together with pattern abce, wereconsidered to represent serologic groups. Theremaining international type strains showedonly minor differences, no more than shown by anumber of random strains. When the additionaltype sera (h, i, k) and examination of autoclavedbacteria were included, a better differentiationof strains was obtained. It was concluded that alimited number of serologic groups exist, whichcan be further differentiated into a large numberof patterns or types. Further investigationsshowed that blocking and 0 inagglutinabilityincrease with the duration of incubation and attemperatures below 37 C (Oeding (72)). In 5-hr-old agar cultures the maximal number of antigensand the strongest agglutinations were obtained.The typing of S. aureus is therefore now donewith live 5-hr-old bacteria incubated at 37 C.The presence of an i antigen is checked in 18-hr-old cultures, as this antigen may develop slowly.Typing of autoclaved bacteria is not performedexcept in spontaneously agglutinating strains,which are usually typable after autoclaving.Recent investigations in this laboratory have,however, shown that in certain instances anantigen may be discovered only after autoclav-ing.Grun (29) and Grun and Kuhn (31) considered

the use of 5-hr-old cultures for agglutination agreat advantage. This modification increasedthe number of antigens recorded and reducedthe number of nontypable strains.

Oeding's serologic classification has proved tobe of definite value in epidemiologic investiga-tions (75, 77, 102). Brodie et al. (7, 8) used thetype strains of Hobbs (36) for the production offactor sera, but Oeding's nomenclature. Immu-nization was performed with bacteria grown on 8

per cent agar. The modifications introduced byBrodie et al. may explain certain discrepancieswith regard to Oeding's results. Nevertheless,very satisfying results were found in epidemio-logic investigations. L6fkvist (57) and Vischer(110) both examined strains from epidemics ofmastitis serologically. Though Oeding's prin-ciple was followed, they used their own strainsfor the production of type sera. L6fkvist foundthe serologic method of value, whereas Vischer'sresults were less promising.

S. Technical problems. It is a general experiencethat adequate techniques are essential for suc-cessful serologic typing of staphylococci. Aconcluding review of the main technical prob-lems will therefore be necessary.

Strains: It is important that well characterizedtype strains be used for serologic investigationsof staphylococci, otherwise results will be diffi-cult to compare. The strains should be kept undercontrolled conditions. In this laboratory monthlysubcultures are made on agar stabs. If any doubtshould arise as to the identity or characteristicsof a strain, it is replaced with a lyophilized cul-ture.

There seems to be a common view that staphy-lococci are antigenically very mutable. This isnot our experience. Although one or two of ourtype strains have shown certain antigenic changesin the course of several years in the laboratory,all other strains have kept their characteristicsremarkably well. Pillet et al. (61, 87) foundCowan's 3 type strains to be stable, and apartfrom two strains which lost their agglutinabilitybut recovered it after passage in rabbits, theirown strains showed no change in antigeniccharacteristics. Grun (25) and Grun and Kuhn(31) found that strains did not change theirantigenic patterns over a longer period. Variantswhich differed from the mother strains in pig-ment, hemolytic activity, or sensitivity to anti-biotics (105) retained their serologic types.Provided the type strains are adequately con-trolled, these observations should insure a soundbasis for serologic classification.

Rabbit immune sera: Cowan produced rabbitimmune sera by the injection of boiled vaccines,thus obtaining antibodies against the heat-stableantigens exclusively. In spite of this simplifica-tion, a satisfactory typing scheme was not ob-tained. A classification which also includes heat-labile antigens (Oeding (66)) is not only possible

3871960]


http://mm

br.asm.org/

Dow

nloaded from


PER OEDING

but in our opinion offers better possibilitiesof a successful result. Immunization shouldtherefore be made with formalin-killed vaccines.

Agglutinin absorption: This should be done withlive or formalin-killed bacteria, which giveidentical results. It is very important that suit-able strains are used for absorption. Pillet andOrta (87) used Cowan's type strain I for absorp-tion of all sera except the homologous one.Christie and Keogh (10) and Hobbs (36) em-phasized the difficulties of absorption in thattoo large an absorbing dose of bacteria tendedto remove specific as well as nonspecific aggluti-nins.

These experiments were made with heat-killed bacteria. When formalin-killed or livebacteria are used for absorption, similar diffi-culties are demonstrated. Successful typing ofS. aureus cannot be achieved unless sufficientlystrong, monovalent sera are obtained. The tech-nique used for absorption is therefore very im-portant.

Pillet et al. (85) used quantitative absorptionwith a defined weight of boiled, dried bacteriafor a given amount of serum. According to theseauthors, absorption is not proportional to theamount of agglutinogen in a strain.The standard technique of Oeding consists of

the absorption of 3 ml serum diluted 1 to 10,with the growth from one Roux bottle of a cer-tain size. This rather empirical technique hasworked quite satisfactorily, although absorptionhas sometimes to be repeated. Brodie (7) pre-ferred repeated absorptions with small portionsof bacteria for 30 min each. A similar techniquehas lately been used by us with good results.Sera are absorbed with defined amounts of wetbacteria in small portions and the titer of spe-cific antibody determined after each absorption.Absorption is considered complete when thespecific titer is not significantly reduced com-pared to the last absorption. As this techniqueis rather time consuming, a large amount ofserum should be produced each time.As pointed out above, "normal" antibodies

will hardly be present in absorbed sera. Itshould therefore not be necessary to use serumdilutions of 1 to 50 or 1 to 100, as done by Mer-cier et al. (61) or Brodie et al. (8). At least, weakantisera should not be diluted more than 1 to 10.

Oeding (66) maintained that a serologic classi-fication of staphylococci should be based upon

an antigenic scheme. The use of monovalentfactor sera containing defined antibodies willthus make it possible to determine the patternof the surface agglutinogens present in a strain.

Agglutination tests: Tests should be made onslides which, in contrast to tube agglutination,give clear-cut reactions. Cowan (13) used boiledbacteria, whereas live 5-hr-old cultures are usedby Oeding (72). Pillet et al. (84) observed thatthe composition of the culture medium on whichbacteria were grown for agglutination influencedthe agglutination titer. In a poor medium littleagglutinogen developed. Grun (26) could not con-firm this observation, and cultivation on nutrientagar can be recommended (Oeding (66)).

According to the technique given by Oeding(66), the slides are agitated for some time andthen examined with the naked eye. Grin (25)prefers to agglutinate in a moist chamber for2 hr and to read with a hand lens. We feel thatthere is a risk of including nonspecific reactionsif the reading is too precise.

In our experience any special procedure toavoid spontaneous agglutination is unnecessaryas the number of unstable strains is small. Whenstrains showing spontaneous agglutination inlive condition are autoclaved, the heat-stableantigens can usually be determined.

B. Epidemiologic Value

In the following the value of serologic methodsfor the typing of S. aureus will be analyzed. Con-sideration will be given both to the results ob-tained with the method described by Cowan,and to those obtained with Oeding's method. Thereports of various authors are, however, notalways comparable, due to the use of differenttyping strains and to technical modifications.Regarding the method of Oeding, his results areonly directly comparable to those of Grun. Forvarious reasons special attention will be paid tothe reports of these two authors and their co-workers.

1. Reproducibility. If a typing system is to beof any use, the results must be reproducible.The antigenic equipment of the staphylococcalsurface is very complex and it appears that thesurface is a somewhat whimsical basis for classi-fication. The technical conditions have beendiscussed above. Here it will only be repeatedthat agglutination of very young staphylococcioffers a great advantage over earlier procedures

-388 [VOL. 24


http://mm

br.asm.org/

Dow

nloaded from



in the effort to demonstrate the total set ofagglutinogens. By this procedure strong antigensare easily demonstrated, whereas weak antigensmay represent an uncertainty. The determina-tion of the antigenic pattern is not so difficultin an epidemiologic investigation. The epidemicstrain may have one weak antigen, which is indi-cated by parentheses, e.g., the antigenic patternabc(h). If a strain has the pattern abc and isconsidered on epidemiologic evaluation to be theepidemic strain, the serologic results should beconsidered affirmative. Reexamination of thestrain will usually disclose the weak h antigen.In the comparison of strains from different mate-rials, e.g., in investigations on frequency dis-tribution, weak antigens certainly represent adifficulty. Similar problems are present in phagetyping, and it is a matter of experience to deter-mine a convenient limit of variation.The reproducibility of serologic typing can be

tested either by repeated examinations on anumber of colonies or subcultures of one strain,or by the examination of strains derived fromthe same source of infection. In investigationsof this kind, not only the technical reproducibil-ity but also the stability of a strain must beconsidered (see above), as well as the possibilityof double infection or contamination.

Gruin (25, 29) and Grun and Kuhn (31),examining the reproducibility of serologic typingby Oeding's method in subcultures of strains andin variants, found the serologic pattern to beunchanged. Grin points out that the possibilityof contamination or double infection shouldbe considered when unexpected results of sero-logic typing are found in laboratory strains or onepidemiologic examination. The antigenic pat-tern is considered a stable characteristic andOeding's serologic typing has given reproducibleresults.A limited number of field investigations has

shown that this serologic method gives as stableresults as can reasonably be expected (27, 51,75, 77, 102). When strains from a common sou-ceare examined, the antigenic patterns are iden-tical or show only minor differences possiblyascribable to the uncertainty connected with aweak antigen present. The evaluation of theidentity of the strains examined was in additionsupported by epidemiologic data, antibiograms,and phage typing.A number of sets of strains have been ex-

amined that were isolated from different siteson one person, from different people, or frompeople and infected fomites in circumstancesthat made it very probable that all the strainshad a common parent. As examples of sets ofstrains which gave completely identical antigenicpatterns can be mentioned 11 strains isolatedfrom different portions of a potato salad respon-sible for food poisoning (75) and 16 independentstrains from 3 outbreaks of food poisoning (77).Kikuth and Grun (51) examining strains ofstaphylococci isolated from various organs ofpatients who had died of staphylococcal infec-tions, found identical antigenic pattern instrains from the same patient.

2. Field examinations. In addition to thereproducibility of a typing system, other proper-ties must be present if it is to be of value inepidemiologic investigations. The number ofpatterns, types, or groups which a method isable to distinguish must not be too small or thecertainty with which an epidemic strain can bepicked out in patients and carriers will suffer.On the other hand, it should be considered adisadvantage if a system registers a great numberof patterns unless they are well characterizedand not influenced by technical conditions.Furthermore, the number of untypable strainsmust not be too large. A method must not be toocomplicated or it will be of limited practicalvalue.By the extension of the original 3 serologic

groups of Cowan (13) to 9 (10, 24) and later 13types (36), a serviceable basis for epidemiologicregistration was established. The method ofCowan has been used with success in field in-vestigations on cutaneous infections (1, 21, 24,37) and on various staphylococcal infections(36); however, Andersen (2) was unable to char-acterize the strains responsible for epidemics ofpemphigus neonatorum. Cross-reactions be-tween some of the types were, however, a diffi-culty in classification (36) and too many strainswere untypable. In two series, Hobbs (36) wasunable to classify 26 and 20 per cent of the strains,respectively, although the number of untypablestrains was reduced by the production of addi-tional sera. The method of Cowan has not beenwidely adopted for epidemiologic work.With the method of Oeding, strains are charac-

terized by their antigenic patterns, which give abetter differentiation of strains than a system

3891960]


http://mm

br.asm.org/

Dow

nloaded from


PER OEDING

which is bound to a limited number of types. Anepidemic strain can be further characterizedby the quantitative appearance of its antigens.In an investigation into the frequency distribu-tion of 239 independent strains, Oeding and Wil-liams (77) found a large number of patterns,although when quantitative variations were nottaken into consideration, four of the patternsaccounted for 50 per cent of the strains. Similarfindings have been made by Grun (29). Themethod will therefore insure a fairly reliablecharacterization of an epidemic strain, but shoulda widely distributed strain be responsible, thepossibility of getting a clear picture of the epi-demic will, as in all sero-diagnostic work, bereduced.Attempts to collect strains with related anti-

genic patterns into types and groups (66, 77)have shown that such a division is mainly ofacademic interest and should not be applied toepidemiologic typing. It has also been shownthat the exclusive use of "group" factor sera(a, b, c, e) (25, 29, 31, 69, 70) does not allow asufficient number of patterns to be distinguished.Therefore when serology is used alone for theclassification of staphylococci, the complete setof factor sera is needed.The number of strains untypable, either be-

cause they have none of the antigens representedin the factor sera or because of spontaneousagglutination, is very small with Oeding'smethod. About 5 per cent of the strains werefound to be untypable by Oeding (70) andGrun (25). When 5-hr-old cultures of 239 inde-pendent strains were examined, 19 were un-typable, of which the total of 11 spontaneouslyagglutinating strains could be typed after auto-claving, i.e., only 3 per cent of the strains couldnot be typed (77). There seems to be a higherproportion of untypable strains from the air,skin, and feces than from pathological material(27, 31).Comparison of serologic typing, phage typing,

and antibiograms has shown that the best re-sults are obtained when two or three of thesemethods are used simultaneously (75, 77, 102).The use of both phage and serologic methods forroutine typing would, however, be too timeconsuming. A laboratory has to choose one ofthem, which can easily be combined with ex-amination of the antibiogram. Grun (29) pre-fers to screen a material by means of antibio-

gram, after which the suspected strains areexamined serologically by means of "group"sera (a, b, c, e) and only when there is still doubtas to the identity, by means of "type" sera (h,i, k, 1). Brodie et al. (7, 8) have used the combinedcharacterization of strains by antibiogram andserological type with success.As mentioned above, we are of the opinion

that as a rule a serologic investigation shouldbe made with the complete set of sera. When anepidemic strain is sufficiently characterized, asimplified serologic screening may be possiblefor further examinations. Recently, however,new antigens have been described by Grun (28)(1) and Haukenes and Oeding (33) (m, n) andinvestigations in progress in this laboratoryindicate the existence of still more antigens. Atyping set of 10 or more factor sera will be toocumbersome for routine examinations. The pro-cedure can be simplified after the principle ofGrun (see above), or screening can be done withpools of factor sera. We feel convinced that thedescription of new antigens in S. aureus willeventually improve serologic classification andexplain observations which are difficult to under-stand today.

S. Correlation to disease. There is no convinc-ing report of correlation between serologic typeand type of infection or virulence in staphylococci(29). Andersen (2) isolated a well characterizedstrain of staphylococci from cases of pemphigusneonatorum in Denmark, and Oeding (65)found a special serologic type responsible for alarge number of cases of mastitis in Norwegianhospitals. The penicillin-resistant mastitis sta-phylococcus was considered to be particularlyvirulent and to have a special affinity for thelactating mammary gland. However, thesestaphylococci have also been found in other in-fections, and pemphigus and mastitis have beencaused by other serologic types of staphylococci.Experience in recent years has shown that a cer-tain strain can dominate a locality for some timeand eventually be replaced by another strain.

4. Correlation to antibiotic sensitivity. Oeding(66, 70) found that strains having the antigenicpattern abc were more frequently resistant topenicillin than strains belonging to other sero-logic types. This was observed in different Nor-wegian hospitals. Pillet et al. (82) observed thatstrains belonging to serologic type 8 were allresistant to chlortetracycline and oxytetracycline.

390 [VOL. 24


http://mm

br.asm.org/

Dow

nloaded from



These authors (66, 70, 82) found it possible thatcertain serologic types are particularly likely todevelop resistance to antibiotics. When oneserologic type shows a higher or lower incidenceof antibiotic resistance, this may naturally beconnected with the frequency with which strainsof this type have been in contact with antibioticsand with the ability of the type to spread in alocality (30, 70). A possible correlation is verydifficult to determine and must await furtherstudy.

5. Correlation to phage types and groups. A fewcomparisons of serologic and phage typing havebeen published, intending to seek a correlationbetween particular types and groups. Hobbs (36)and Wahl and Fouace (111) found that therewas a tendency for strains of Cowan's types Ito III to fall into the later established phagegroups I to III. Oeding (69) observed thatstrains with serologic pattern abc (Cowan'stype III) usually belonged to phage group III.Oeding and Vogelsang (76) and Pillet et al.(82) found good correlation between serologicpattern abe (Cowan's group I) and phage groupI, although strains belonging to this serologicgroup were frequently not typable by phages(36, 76).Oeding and Williams (77) found that all

strains lysed by phage 187 possessed the kantigen. Phage 3A strains were often spon-taneously agglutinable; such strains in an earlierinvestigation (76) were frequently connectedwith phage 70.

It can be concluded from these reports thatthere is a broad correlation between phage typ-ing and serologic groups, but only a certaintendency for association between types deter-mined by the two methods. In particular there isno example of complete correlation and in somecases the reactions are widely spread (77).Whereas Pillet et al. (82) believe the receptors

for agglutinins to be associated with the re-ceptors for phages, Oeding and Williams (77)conclude that they are different.

IV. CONCLUSIONIt is mainly because hospital infections caused

by Staphylococcus aureus have developed into aserious problem during recent years that therehas been an increasing interest in this organism.A vast number of publications have appearedwhich deal with the epidemiologic conditions in

staphylococcal infection. It is undoubtedly veryimportant that we should learn the routes ofinfection and the conditions that exist in ourmodern hospitals where such infections occur.Although the yellow staphylococcus has alwaysbeen one of our most important pathogens, it is,however, astonishing how little progress hasbeen made during the last decades in basic re-search on this organism. We feel that moreknowledge of the fundamental characteristics ofS. aureus will help us to understand and combatstaphylococcal epidemics. It is our impressionthat this view is today shared by a number ofmicrobiologists who have taken up staphylo-coccal research on an experimental basis.

For a long time the antigenic properties of S.aureus have been considered very complex andpeople have therefore felt reluctant to startwork in this field. Immunologic and serologicwork on staphylococci does certainly presentmany difficulties but this should stimulaterather than impede the taking up of these im-portant and topical problems. The literature onthe antigenic properties of S. aureus is scatteredand extends over a long period of time. Ourknowledge is so limited that one might wellsay that a critical review should wait. Never-theless it was felt that at the present time a dis-cussion of the problems might be a help to thoseworking with or planning to work with staphy-lococcal immunology and serology.Many gaps have to be filled before we shall

be able to get a clear picture of the immunologicand serologic faculties of S. aureus. It is thereforedifficult to say on which points it may be mostadvantageous to concentrate research at present.We feel that a registration of the antigens of S.aureus, their nature, and their properties shouldhave high priority. As the basic research mustbe done on a few selected strains it is necessary,if we are to be able to compare the results, forinternational type strains to be used. Theseshould be kept under controlled conditions toavoid variations in their characteristics. Fur-thermore, the results obtained with differentantigen-antibody reactions should be correlatedand the nomenclature of the antigens madeuniform. Increased knowledge of the antigens ofS. aureus will eventually improve serologic typeclassification and facilitate work in other fieldsof staphylococcal'research.

3911960]


http://mm

br.asm.org/

Dow

nloaded from


PER OEDING

V. REFERENCES

1. ALLISON, V. D. AND HOBBS, B. C. 1947 Aninquiry into the epidemiology of pemphigusneonatorum. Brit. Med. J., 2, 1-7.

2. ANDERSEN, E. K. 1943 Studies on thespecificity of pemphigus staphylococci.Acta Pathol. Microbiol. Scand., 20, 242-256.

3. ANDERSEN, E. K. AND HEILESEN, B. 1951On the occurrence and typing of Staphy-lococcus pyogenes in weeping exzema. ActaDermato-Venereol., 31, 679-703.

4. BEiSER, S. M., DWORETZKY, M., SMART,K. M., AND BALDWIN, H. S. 1958 Pre-cipitating antibodies to staphylococcalantigens in human sera. J. Allergy, 29,44-47.

5. BLAIR, J. E. 1958 Factors determining thepathogenicity of staphylococci. Ann. Rev.Microbiol., 12, 491-506.

6. BOOR, A. K. AND MILLER, C. P. 1934 Astudy on bacterial proteins with specialconsideration of gonococcus and meningo-coccus. J. Exptl. Med., 59, 63-74.

7. BRODIE, J. 1957 Some observations on theserological typing of Staphylococcus pyo-genes. J. Clin. Pathol., 10, 215-218.

8. BRODIE, J., JAMIESON, W., AND SOMMER-VILLE, T. 1955 Complications of oxytet-racycline and tetracycline therapy relatedto a defined type of resistant Staphylococcus.Lancet, 2, 223-225.

9. CHRISTIE, R. 1940 The use of slide ag-glutination to determine pathogenicity ofstaphylococci. Australian J. Exptl. Biol.Med. Sci., 18, 397-400.

10. CHRISTIE, R. AND KEOGH, E. V. 1940 Phys-iological and serological characteristicsof staphylococci of human origin. J.Pathol. Bacteriol., 51, 189-197.

11. COHEN, S. G., DZURY, D. S., AND MICHELINI,F. J. 1958 Immune responses to Staphy-lococcus aureus antigens. J. InfectiousDiseases, 103, 61-66.

12. COWAN, S. T. 1938 The classification ofstaphylococci by precipitation and bio-logical reactions. J. Pathol. Bacteriol.,46, 31-45.

13. COWAN, S. T. 1939 Classification ofstaphylococci by slide agglutination. J.Pathol. Bacteriol., 48, 169-173.

14. CUMMINS, C. S. AND HARRIS, H. 1956 Thechemical composition of the cell wall insome Gram-positive bacteria and its possi-ble value as a taxonomic character. J.Gen. Microbiol., 14, 583-400.

15. DUDGEON, L. S. AND BAMFORTH, J. 1925

On staphylococcal precipitin reactions incases of acute and chronic infections andalso in serum sickness. J. Hyg., 23, 375-388.

16. DWORETZKY, M., BALDWIN, H. S., ANDSMART, K. M. 1956 Biologic effects ofmaterials derived from staphylococcus:Further studies. J. Allergy, 27, 39-49.

17. DWORETZKY, M., BEISER, S. M., AND BALD-WIN, H. S. 1958 Immunologic relation-ship between a nucleoprotein fraction oftwo strains of staphylococcus as deter-mined by anaphylaxis. J. Allergy, 29,48-55.

18. DWORETZKY, M., ZEITLIN, B. R., KAHN,M. C., AND BALDWIN, H. S. 1952 Phar-macologic and anaphylactogenic propertiesof staphylococcus protein and unfrac-tionated extract. J. Allergy, 23, 359-364.

19. DYAR, M. R. 1958 Electrokinetical studieson bacterial surfaces. II. Studies on surfacelipids, amphoteric material, and someother surface properties. J. Bacteriol.,56, 821-834.

20. ELEK, S. D. 1959 Staphylococcus pyogenesand its relation to disease. E. & S. Living-stone, Ltd., Edinburgh and London.

21. ELLIOT, S. D., GILLESPIE, E. H., AND HOL-LAND, E. 1941 An outbreak of "pemphi-gus neonatorum" in a maternity home.Lancet, 1, 169-171.

22. FELLOWES, 0. N. AND ROUTH, J. I. 1944 Astudy of the chemical composition andantigenic properties of a polysaccharidefraction of three-hour cultures of Staphy-lococcus aureus. J. Lab. Clin. Med., 29,1054-1061.

23. FRASER, K. B. 1954 Antibody in glandular-fever sera to an antigen common to strep-tococci and staphylococci. J. Pathol.Bacteriol., 67, 301-309.

24. GILLESPIE, E. H., DEVENISH, E. A., ANDCOWAN, S. T. 1939 Pathogenic staphy-lococci; their incidence in the nose and onthe skin. Lancet, 2, 870-873.

25. GRUTN, L. 1957 Serologische Untersuchun-gen an pyogenen Staphylokokken. Z.Hyg. Infektionskrankh., 144, 238-247.

26. GRUN, L. 1958 Alte und neue Problemeder Staphylokokken. Ergeb. Mikrobiol.Immunitatsforsch. u. Exptl. Therap., 31,129-183.

27. GRUN, L. 1958 Untersuchungen an Darm-staphylokokken. Arch. Hyg., 142, 3-7.

28. GRUN, L. 1958 tUber ein neues Staphylo-kokken-Antigen. Z. Hyg. Infektionskrankh.,145, 259-262.

29. GRUN, L. 1959 Die serologische Differ-

392 tVOL. 24


http://mm

br.asm.org/

Dow

nloaded from



enzierung von Staphylokokken. Z. Hyg.Infektionskrankh., 146, 129-141.

30. GRUN, L. 1959 tUberBeziehungenzwischenPigmentation, Antibiogramm und Anti-genstruktur bei pathogenen Staphylokok-ken. Zentr. Bakteriol. Parasitenk., 174,184-191.

31. GRtN, L. AND KUHN, H. 1958 Erfahrungenund Ergebnisse bei der serologischenUntersuchung von Staphylococcus pyogenesaureus. Z. Hyg. Infektionskrankh., 144,535-548.

32. HARRIS, J. 0. 1953 Electrophoretic be-havior and crystal violet adsorption capac-ity of ribonuclease treated bacterial cells.J. Bacteriol., 65, 518-521.

33. HAUKENES, G. AND OEDING, P. 1960 Ontwo new antigens in Staphylococcus aureus.Acta Pathol. Microbiol. Scand., 49, 237-248.

34. HAYES, L. 1951 Specific serum agglutina-tion of sheep erythrocytes sensitized withbacterial polysaccharides. Australian J.Exptl. Biol. Med. Sci., 29, 51-62.

35. HINE, T. G. M. 1922 Serological classifica-tion of the staphylococci. Lancet, 2,1380-1382.

36. HoBBs, B. C. 1948 A study ot the serolog-ical type differentiation of Staphylococcuspyogenes. J. Hyg., 46, 222-238.

37. HOBBs, B. C., CARRUTHERS, H. L., ANDGOUGH, J. 1947 Sycosis barbae. Serolog-ical types of Staphylococcus pyogenes innose and skin and results of penicillintreatment. Lancet, 2, 572-574.

38. HOFFSTADT, R. E. AND CLARK, W. M. 1935The carbohydrates of the rough and smoothforms of Staphylococcus aureus. J. In-fectious Diseases, 56, 250-254.

39. ILLCHMANN-CHRIST, A. AND NAGEL, V. 1954Ein Beitrag zur serologischen Identi-fizierung bakterieller blutgruppenspezi-fischer Antigendeterminanten durch dieAbsorption menschlicher Iso-Normalseren.Z. Immunitatsforsch., 111, 307-333.

40. INOUE, S. 1942 Studien uber die giftigenLeibessubstanzen von Staphylococcusaureus. T6hoku J. Exptl. Med., 43, 1-12.

41. ISHIMOTO, N., SAITO, M., AND ITO, E. 1958Composition of the cell wall of Staphylo-coccus aureus 209 P. Nature, 182, 959-960.

42. JENSEN, K. 1958 A normally occurringstaphylococcus antibody in human serum.Acta Pathol. Microbiol. Scand., 44, 421-428.

43. JENSEN, K. 1959 Unders6gelser overstafylococcernes antigenstruktur. Thesis,E. Munksgaard, Copenhagen.

44. JULIANELLE, L. A. 1922 Studies of hemo-lytic staphylococci. J. Infectious Dis-eases, 31, 256-284.

45. JULIANELLE, L. A. AND HARTMANN, A. F.1936 The immunological specificity ofstaphylococci. IV. Cutaneous reactionsto the type-specific carbohydrates. J.Exptl. Med., 64, 149-159.

46. JULIANELLE, L. A. AND WIEGHARD, C. W.1934 Immunological specificity of car-bohydrates derived from staphylococci.Proc. Soc. Exptl. Biol. Med., 31, 947-949.

47. JULIANELLE, L. A. AND WIEGHARD, C. W.1935 The immunological specificity ofstaphylococci. I. The occurrence of serolog-ical types. J. Exptl. Med., 62, 11-21.

48. JULIANELLE, L. A. AND WIEGHARD, C. W.1935 The immunological specificity ofstaphylococci. III. Interrelationships ofcell constituents. J. Exptl. Med., 62,31-37.

49. KAHN, M. C., BALDWIN, H. S., ZEITLIN,B. R., AND SMART, M. 1951 Cutaneousreactions to staphylococcus polysac-charide, protein, and an unfractionatedextract in hypersensitive and normalindividuals. J. Allergy, 22, 237-248.

150. KEOGH, E. V., NORTH, E. A., AND WAR-BURTON, M. F. 1948 Adsorption ofbacterial polysaccharides to erythrocytes.Nature, 161, 687-688.

51. KIKUTH, W. AND GRUN, L. 1957 ZurHygiene und Bakteriologie des Stafy-lokokken-Hospitalismus. Deut. med.Wochschr., 82, 549-553.

52. KLIENEBERGER-NOBEL, E. 1948 Capsulesand mucoid envelopes of bacteria. J.Hyg., 46, 345-348.

53. KOLLE, W. AND OTTO, R. 1902 Die Differ-encirung der Staphylokokken mittelst derAgglutination. Z. Hyg. Infektions-krankh., 41, 369-379.

54. KoURILSsY, R., PIERON, R., AND RICHOU,R. 1957 Le comportement diff6rent desh6maties humaines et animales dans lareaction d'immuno-adh6rence vis-a-visdu staphylocoque. Compt. rend., 244,957-959.

55. LANCEFIELD, R. C. 1925 The immunolog-ical relationships of Streptococcus viridansand certain of its chemical fractions. II.Serological reactions obtained with anti-nucleoprotein sera. J. Exptl. Med., 42,397-412.

56. LYONS, C. 1937 Antibacterial immunityto Staphylococcus pyogenes. Brit. J. Exptl.Pathol., 18, 411-422.

1960] 393


http://mm

br.asm.org/

Dow

nloaded from


PER OEDING

57. LOFEVIST, TH. 1957 Serologic identifica-tion of coagulase-positive staphylococcifrom outbreaks of mastitis. Acta Pathol.Microbiol. Scand., 41, 521-536.

58. MCCARTY, M. 1959 The occurrence ofpolyglycerophosphate as an antigeniccomponent of various gram-positive bac-terial species. J. Exptl. Med., 109, 361-378.

59. MADSEN, S. T. 1956 Studies on the ag-

glutination of red cells sensitized withbacterial extracts. Thesis, Univers. iBergen, med. rekke nr. 4.

60. MAGGI, N. 1938 Sulla specificitA deipolisaccaridi degli stafilococchi. Riv.patol. sper., 21, 261-278.

61. MERCIER, P., PILLET, J., AND CHABANIER,P. 1950 Determination des types destaphylocoques par l'agglutination. Ann.inst. Pasteur, 78, 457-466.

62. MITCHELL, P. AND MOYLE, J. 1951 Isola-tion of hydrolytic products of a glycero-phospho-compound from Micrococcus pyo-

genes. J. Gen. Microbiol., 5, 966-980.63. MITCHELL, P. AND MOYLE, J. 1951 The

glycerophospho-protein complex envelopeof Micrococcus pyogenes. J. Gen. Micro-biol., 5, 981-992.

64. NETER, E. AND GORZYNSKI, E. A. 1959Studies on indirect staphylococcal hemag-glutination. Z. Immunitatsforsch., 118,269-288.

65. OEDING, P. 1952 Examinations on peni-cillin resistant, serologically homogeneousstaphylococci isolated from human masti-tis. Acta Pathol. Microbiol. Scand., 31,145-163.

66. OEDING, P. 1952 Serological typing ofstaphylococci. Acta Pathol. Microbiol.Scand., Suppl. 93, 356-363.

67. OEDING, P. 1953 A simple serologicalprocedure for the differentiation betweenpyogenic and saprophytic staphylococci.Acta Pathol. Microbiol. Scand., 32, 187-192.

68. OEDING, P. 1953 Serological typing ofstaphylococci. II. Investigations on theserological, chemical and physical prop-

erties of the antigens. Acta Pathol.Microbiol. Scand., 33, 312-323.

69. OEDING, P. 1953 Serological typing ofstaphylococci. III. Further investigationsand comparison to phage typing. ActaPathol. Microbiol. Scand., 33, 324-336.

70. OEDING, P. 1954 Staphylococcal studiesin hospital staffs. IV. Serological investi-gation. Acta Pathol. Microbiol. Scand.,34, 34-46.

71. OEDING, P. 1954 Immunochemical studiesof staphylococcal polysaccharides. I.Methods for extraction and purification.Acta Pathol. Microbiol. Scand., 35, 484-494.

72. OEDING, P. 1957 Agglutinability of pyo-genic staphylococci at various conditions.Acta Pathol. Microbiol. Scand., 41, 310-324.

73. OEDING, P. 1957 Observations on thesensitization of erythrocytes with staphy-lococcal polysaccharide products. ActaPathol. Microbiol. Scand., 41, 435-448.

74. OEDING, P. 1959 Staphylococcal antigen-antibody reactions in agar. Acta Pathol.Microbiol. Scand., 47, 53-64.

75. OEDING, P. AND SOMPOLINSKY, D. 1958On typing of Staphylococcus aureus: Therelationship between typing according tophage patterns, serology, and antibio-grams. J. Infectious Diseases, 102, 23-34.

76. OEDING, P. AND VOGELSANG, TH. M. 1954Staphylococcal studies in hospital staffs.V. Comparison between serological typingand phage typing. Acta Pathol. Micro-biol. Scand., 34, 47-56.

77. OEDING, P. AND WILLIAMS, R. E. 0. 1958The type classification of Staphylococcusaureus: A comparison of phage-typingwith serological typing. J. Hyg., 56,445-454.

78. PAKULA, R. AND WALCZAK, W. 1956 Unfacteur sensibilisant les globules rouges,commun aux staphylocoques et aux strep-tocoques h6molytiques. Ann. inst. Pas-teur, 91, Suppl. 54-61.

79. PARK, J. T. AND STROMINGER, J. L. 1957Biochemical basis for the mechanism ofaction of penicillin and for its selectivetoxicity. Science, 125, 99-101.

80. PATTISON, I. H. AND MATTHEWS, P. R. J.1957 Observations on the serologicalidentification of bovine coagulase-positivestaphylococci. J. Pathol. Bacteriol., 74,335-346.

81. PENNER, M. AND VOLDRICH, M. 1932Staphylococcus aureus agglutinins in tuber-culous effusions. J. Infectious Diseases,50, 143-145.

82. PILLET, J., CALMELS, J., ORTA, B., ANDCHABANIER, G. 1954 Etude compardedu type s6rologique de la sensibilit6 auxbact6riophages et de l'antibiogramme de201 souches de staphylocoques isol6s parprelevements syst6matiques chez le nour-risson. Ann. inst. Pasteur, 86, 309-319.

83. PILLET, J., ISBIR, S., AND MERCIER, P. 1950

394 [VOL. 24


http://mm

br.asm.org/

Dow

nloaded from



Etude biologique des staphylocoquesd'origine animale. Ann. inst. Pasteur,78, 638-643.

84. PILLET, J., MERCIER, P., AND ORTA, B. 1951Influence du milieu de culture sur laproduction des agglutinogenes staphy-lococciques. Ann. inst. Pasteur, 81, 224-227.

85. PILLET, J., MERCIER, P., AND ORTA, B. 1952Recherches sur les agglutinines hdt6rolo-gues des serums antistaphylococciques. I.Utilasion d'une m6thode d'absorbtionquantitative.Ann. inst. Pasteur, 82, 465-470.

86. PILLET, J. AND ORTA, B. 1952 Recherchessur les agglutinines h6t6rologues des ser-ums antistaphylococciques. II. Etude del'agglutination par les s6rums bruts. Ann.inst. Pasteur, 82, 488-492.

87. PILLET, J. AND ORTA, B. 1953 Recherchessur les agglutinines het6rologues dess6rums antistaphylococciques. III. 1Ptudede la sp6cificit6 des types 4, 5, 6, 7, 8, 9.Ann. inst. Pasteur, 84, 420-426.

88. PILLET, J. AND ORTA, B. 1954 Recherchessur la composition antig6nique des staphy-locoques non pathogenes. I. Mise en evi-dence d'agglutinogenes specifiques etd'un agglutinogene commun aux souchespathogenes et non pathogenes. Ann.inst. Pasteur, 86, 752-758.

89. PILLET, J., ROUYER, M., AND ORTA, B. 1955Recherches sur la nature chimique desagglutinogenes staphylococciques. Ann.inst. Pasteur, 88, 662-665.

90. PRICE, K. M. AND KNEELAND, Y., JR. 1954A mucoid form of Micrococcus pyogenesvar. aureus which shows capsular swellingwith specific immune serum. J. Bacteriol.,67, 472-475.

91. PRICE, K. M. AND KNEELAND, Y., JR. 1956Further studies of the phenomenon ofcapsular swelling of Micrococcus pyogenesvar. aureus in the presence of immuneserum. J. Bacteriol., 71, 229-230.

92. RANTZ, L. A., RANDALL, E., AND ZUCKERMAN,A. 1956 Hemolysis and hemagglutina-tion by normal and immune serums oferythrocytes treated with a nonspeciesspecific bacterial substance. J. Infec-tious Diseases, 98, 211-222.

93. RANTZ, L. A., ZUCKERMAN, A., AND RANDALL,E. 1952 Hemolysis of red blood cellstreated by bacterial filtrates in the pres-ence of serum and complement. J. Lab.Clin. Med., 39, 443-448.

94. ROUNTREE, PH.M. AND BARBOUR, R. G. H.

1952 Antibody to the erythrocyte-coatingpolysaccharide of staphylococci: Its oc-currence in human sera. AustralasianAnn. Med., 1, 80-83.

95. SACHSE, H. 1956 Untersuchungen uberStreptokokken-Hamosensitine. Z. Immu-nitatsforsch., 113, 333-342.

96. SALTON, M. R. J. 1953 Cell structure andthe enzymic lysis of bacteria. J. Gen.Microbiol., 9, 512-523.

97. SALTON, M. R. J. 1957 Cell-wall amino-acids and amino-sugars. Nature, 180,338-339.

98. SAVINO, E. 1932 Anticorps du serum anti-staphylococcique. Compt. rend. soc. biol.,109, 328-329.

99. SEEDORF, J. 1924 Staphylococci. Thesis,E. Munksgaard, Copenhagen.

100. SEELIGER, H. P. R. AND SULZBACHER, F.1956 Antigenic relationships between Lis-teria monocytogenes and Staphylococcusaureus. Can. J. Microbiol., 2, 220-231.

101. SHRIGLEY, E. W. 1945 Agglutination ofStaphylococcus aureus in the presence ofchorio-allantoic fluid from hens' eggs.Science, 102, 64-65.

102. SOMPOLINSKY, D., HERMANN, Z., OEDING,P., AND RIPPON, J. E. 1957 A series ofpostoperative infections. J. InfectiousDiseases, 100, 1-11.

103. SPINK, W. W. 1939 Attempts to demon-strate a surface antigen of staphylococciand specific phagocytosis. Proc. Soc.Exptl. Biol. Med., 40, 549-552.

104. STACEY, M. 1949 The nature of the surfaceof Gram-positive bacteria. In The natureof the bacterial surface, pp. 29-46. Editedby A. A. Miles and N. W. Pirie. BlackwellScientific Publications, Oxford.

105. STERN, H. AND ELEK, S. D. 1957 Antigenicstructure of Staphylococcus pyogenes. J.Pathol. Bacteriol., 73, 473-483.

106. STRANGE, R. E. AND POWELL, J. F. 1954Hexosamine-containing peptides in sporesof Bacillus subtilis, B. megatherium andB. cereus. Biochem. J., 58, 80-85.

107. THOMPSON, R. AND KHORAZO, D. 1937Correlated antigenic and biochemical prop-erties of staphylococci. J. Bacteriol., 34,69-79.

108. TRIGUEROS, E. A. 1940 Composici6n qui-mica de los estafilococos (Polisac~ridos yproteinas). Rev. clin. espafn., 1, 160-161.

109. VERWEY, W. F. 1940 A type-specific anti-genic protein derived from the staphylo-coccus. J. Exptl. Med., 71, 635-644.

110. VISCHER, D. 1959 Serologische Unter-

19601 395


http://mm

br.asm.org/

Dow

nloaded from


PER OEDING

suchungen an Staphylokokken. Schweiz.Z. allgem. Pathol. u. Bakteriol., 22, 42-62.

111. WAHL, R. AND FOUACE, J. 1952 Sur lesprincipes de classification des staphylo-coques pathogenes par la m6thode desphages. Ann. inst. Pasteur, 82, 542-555.

112. WEBB, M. 1948 The action of lysozyme on

heat-killed Gram-positive microorganisms.J. Gen. Microbiol., 2, 260-274.

113. WIEGHARD, C. W. AND JULIANELLE, L. A.1935 The immunological specificity ofstaphylococci. II. The chemical natureof the soluble specific substances. J.Exptl. Med., 62, 23-30.

114. YONEMURA, N. 1936 Ueber die im-munisatorische Einteilung von Staphy-lococcus pyogenes. Z. Immunit itsforsch.,89, 392-397.

396 [VOL. 24


http://mm

br.asm.org/

Dow

nloaded from


PROPERTIES OF STAPHYLOCOCCUSAUREUS · Molisch test was strongly positive, whereas the biuret test...

Documents

Transcript of PROPERTIES OF STAPHYLOCOCCUSAUREUS · Molisch test was strongly positive, whereas the biuret test...