Degradation studies on Escherichia coli capsular polysaccharides by bacteriophages

FEMS Microbiology Letters 153 (1997) 105^110

Degradation studies on

capsular polysaccharides by bacteriophages Escherichia coli

Wolfgang Nimmich * Institut fu ë r Medizinische Mikrobiologie, Medizinische Fakulta ë t, Universita ë t Rostock, Schillingallee 70, D-18057 Rostock, Germany

Received 25 February 1997; revised 18 April 1997; accepted 20 May 1997

Abstract

The serologically and structurally related Escherichia coli capsular polysaccharides (K antigens) K13, K20, and K23 were found to be depolymerized by the bacteriophages xK13 and xK20 to almost similar oligomer profiles as shown by polyacrylamide gel electrophoresis. The phage-polysaccharide interactions were followed by an increase of reducing 2-keto-3deoxyoctulosonic acid due to a phage-associated glycanase that catalyzed the hydrolytic cleavage of common Lketopyranosidic 2-keto-3-deoxyoctulosonic acid linkages. The related E. coli K antigens K18, K22, and K100 as well as the Haemophilus influenzae type b capsular polysaccharide were degraded by bacteriophage xK100 with different efficacy. It is suggested that xK100 enzymatically cleaves ribitol-5-phosphate bonds as the only structural feature present in all the polysaccharides investigated. Keywords : Escherichia coli

; K-speci¢c bacteriophage; Capsular polysaccharide ; Depolymerase enzyme

1. Introduction

Several bacteriophages are known to speci¢cally recognize Escherichia coli capsular polysaccharides (K antigens) as primary receptors. They were found to be associated with enzymatic activity giving rise to depolymerization of the respective K antigenic polysaccharides [1,2]. Phages harboring glycanase activity towards polysaccharides of the former L group antigens have gained increasing interest since these antigens are associated with E. coli strains causing extraintestinal infections [3,4]. The present study shows the host range of some E. coli K-speci¢c bacteriophages and the phage-associated enzymatic degradation of isolated homologous * Tel.: +49 (381) 494 5911; Fax: +49 (381) 494 5902.

and heterologous capsular polysaccharides. The investigations have been focused on two groups of serologically cross-reacting and chemically related K antigens, K13, K20 and K23 [5] and K18, K22 and K100 [6,7]. Also the Haemophilus in£uenzae type b capsular polysaccharide was included because of its cross-reactivity to E. coli K100 [8].

2. Materials and methods

2.1. Bacterial strains

The E. coli K test strains from the collection of the International Escherichia and Klebsiella Centre, Statens Seruminstitut, Copenhagen, Denmark, were used [6].

0378-1097 / 97 / $17.00 ß 1997 Federation of European Microbiological Societies. Published by Elsevier Science B.V. PII S 0 3 7 8 - 1 0 9 7 ( 9 7 ) 0 0 2 4 2 - 5

FEMSLE 7670 20-10-97

106

W. Nimmich / FEMS Microbiology Letters 153 (1997) 105^110

2.2. Bacteriophages

xK13 has recently been isolated and characterized [9]. xK20 was provided by S. Stirm, Biochemisches Institut der Universitaët GieMen, Germany. A K100speci¢c bacteriophage xK100 was isolated from local sewage using E. coli strain F147 (O75:K100:H5) as host and standard procedures as described previously [2,9]. After puri¢cation by isopycnic centrifugation on a discontinuous cesium chloride gradient the phage suspension had a titer of 2U1012 pfu ml31 . xK100 was kept over chloroform. 2.3. Host range of the phages

Drops of phage suspensions (1U109 pfu ml31 ) were applied onto agar plates with freshly seeded lawns of the E. coli K test strains. After overnight incubation at 37³C the plates were scored for plaques with con£uent lysis. 2.4. Capsular polysaccharides

The E. coli capsular polysaccharides K13, K18, K20, K22, K23, and K100 were a generous gift from B. and K. Jann, Max Planck Institut fuër Immunbiologie, Freiburg, Germany. The Haemophilus in£uenzae type b polysaccharide was kindly supplied by K.-D. Hungerer, Behringwerke AG, Marburg, Germany. 2.5. Polysaccharide depolymerization

Phage-catalyzed depolymerization was carried out as described previously [2,4]. Brie£y, K antigen samples (4 mg ml31 ) were incubated with puri¢ed phage suspensions containing 2U1010 pfu ml31 for 16 h at 37³C. Solutions without phage particles served as control. Specimens were analyzed for reducing groups with the ferricyanide reagent [10]. Ribose and 2-keto-3-deoxyoctulosonic acid (KDO) were used for calibration. Reducing KDO was determined by a modi¢ed periodate/thiobarbituric acid test [2,11]. 2.6. Polyacrylamide gel electrophoresis (PAGE)

Samples of phage-degraded capsular antigens were

subjected to vertical PAGE using 25% polyacrylamide and TBE bu¡er (90 mM Tris-borate, 30 mM EDTA; pH 8.0). Gels were developed by a combined Alcian blue-silver staining method [2,4]. 3. Results

3.1. Group K13, K20 and K23

The results of the host range studies are shown in Table 1. E. coli phage xK13 did not show any lytic activity towards K20 and K23. It was not possible to propagate the phage on these strains. Phage xK20 showed complete lysis of the K20 and K23 test strains but no reaction with strain K13 [9]. The negative reaction is suggested to be due to the presence of restriction systems. Also K5 strains were attacked by xK20 as observed previously [2,4]. The K13, K20, and K23 antigens are serologically related [5]. The basis for the cross-reactivity was found in the chemical structure of the antigens. The three polysaccharides had a disaccharide repeating unit composed of (3-L-Rib-1,7-L-KDOp-2) in common. K13 carries an additional O-acetyl group at C-4 of the KDO. The K20 antigen is O-acetylated at C-5 of the ribose moiety [5]. To obtain more information about phage speci¢city isolated capsular polysaccharides were directly exposed to high concentrations of puri¢ed phages xK13 and xK20. The reactivity of the phages was followed up by measuring the reducing power of the suspension. The kinetics of the K13 and K23 polysaccharide degradations are shown in Fig. 1. xK13 liberated about 300 nmol of reducing KDO from K13 and K23. xK20 proved to be less e¤cient in both heterologous systems with 250 and 270 nmol reducing KDO liberated from K13 and K23, respectively. This corresponds well to the value of the xK20/K20 system previously reported [2]. The formation of depolymerization products was demonstrated by PAGE. The results are shown in Fig. 2. It is evident that xK13 in the homologous K13 and heterologous K23 systems showed similar degradation pro¢les. K20, however, was depolymerized to one oligomer of lower molecular mass. The reaction of phage xK20 on K13 and K23 apparently pro-

FEMSLE 7670 20-10-97

W. Nimmich / FEMS Microbiology Letters 153 (1997) 105^110

107

3

x

x

1 Fig. 1. Liberation of reducing KDO during incubation of the polysaccharides K13 and K23 (4 mg ml ) with phages K13 and K20 10 1 10 pfu ml ). Untreated K13 and K23 served as controls. 50 l samples were analyzed with the ferricyanide reagent at indicated

(1

U

3

W

times.

duced the same main product as in the

xK13/K13

From the results it may be assumed that

xK13 has xK20

system. Three degradation products were obtained

the same substrate speci¢city as reported for

from K20 by

[12] and catalyzed the hydrolysis of

xK20.

Similar results were reported

L-octulopyrano-

by Altmann et al. [12] who studied, however, the

sidonic linkages present in the three polysaccharides

L-ketofuranosidic link-

cleavage of the de-O-acetylated K20 (i.e. the K23

studied. The K95 glycan with

antigen) by this phage and identi¢ed a tetrasacchar-

ages of KDO is not a substrate of the enzymes. The

ide as the main product besides smaller amounts of

presence of

hexa- and octosaccharides [12]. As known from the

the K13 and K20 polysaccharides obviously did not

relevant literature the enzyme-catalyzed degradation

prevent access of the phages.

process essentially stops at the tetrasaccharide stage [1].

O-acetyl

groups in di¡erent positions of

xK20

The lytic activity of

towards K5 strains

could be veri¢ed by depolymerization of the K5

Table 1 Host range of some

E. coli

E. coli

K phages

serovar

Strain No.

Lytic activity (1

xK13 O6 :K13 :H1 O21 :K20 :H

Su4344-41

3

E19a

O25 :K23 :H1

H54

O75 :K100 :H5

F147

O23 :K18 :H15

E39a

O23 :K22 :H15

H67

O10 :K5 :H4

Bi8337-41

O75 :K95 :H5

F3b

+, complete lysis ;

3

+

3 3 3 3 3 3 3

, no lysis.

FEMSLE 7670 20-10-97

U

xK20

3 + +

3 3 3 +

3

9 1 10 pfu ml )

3

xK100

3 3 3 + + +

3 3

xK5

3 3 3 3 3 3 + +

W. Nimmich / FEMS Microbiology Letters 153 (1997) 105^110

108

xK13- and xK20-degraded capsular polysaccharides K13, K20, and K23. xK13. Lane 3, K13 degraded by xK20. Lane 4, K20 degraded by xK20. Lane 6, K20 degraded by xK13. Lane 7, K23 degraded by xK13. Lane 9, K23 degraded by xK20.

Fig. 2. Alcian blue-silver-stained PAGE gel of untreated and

Lanes 2, 5, 8 : untreated K13, K20, and K23. Lane 1, K13 degraded by

polysaccharide shown, the

this

existence

enable

xK20

(not

shown

outstanding of to

two

here). reaction

di¡erent

As

previously

was

due

glycanases

depolymerize

quite di¡erent polysaccharides [2,4].

the

3.2. Group K18, K22, K100, and Hib

to

which

chemically

This group of strains and capsular antigens was studied with

xK100 which was isolated by standard

procedures [2] and found to be useful for the detec-

Fig. 3. Lanes 1, 3, 5, 7 : untreated K22, K100, K18, and Hib, respectively. Lane 2, 4, 6, 8 : spectively.

FEMSLE 7670 20-10-97

xK100-treated K22, K100, K18 and Hib, re-

W. Nimmich / FEMS Microbiology Letters 153 (1997) 105^110

E. coli

tion of

K100 strains from patients with uri-

nary tract infections [13]. Host range studies revealed that

x

109

activity it is tempting to postulate that conformational aspects are to be considered too.

K100 not only

E. coli bacteriophages xK13, xK20

The K-speci¢c

xK100

lysed the homologous K100 strain used as host but

and

also the K18 and K22 test strains out of the collec-

the homologous and related heterologous capsular

tion of the known 77

E. coli

not

these

surprising

since

K test strains. This was strains

were

known

to

cross-react serologically [6]. Another cross-reaction has

been

reported

and

Haemophilus in£uenzae

E. coli

between

studies

of

the

polysaccharides.

Further-

more, such oligosaccharides are important products

to high molecular mass protein carriers. [5]. Puri¢ed

[14].

polysaccharides themselves proved to be only poorly

concerned

the

occurrence

of

of

this

a

on

structural

for the development of vaccines after being coupled

aspect

based

tion of oligomers which thus become available for

the

interesting

(Hib)

polysaccharides. They may be used for the produc-

respective structures of the capsular polysaccharides An

b

O75 :K100 :H5

were found to be able to depolymerize

cross-reactivity

`natural

immunity'

immunogenic

against Hib infections as a result of anti-K100 anti-

phages

bodies [8].

coli

x

The interaction between

K100 and the isolated

capsular polysaccharides K18, K22, K100, and Hib

for

from

in

the

infants

[15].

identi¢cation

extra-intestinal

The of

K

application

of

antigens

E.

infections

has

in

already

been veri¢ed as a convenient and inexpensive method [3,4,13].

as performed by PAGE is demonstrated in Fig. 3. It is

evident

formed. complete was

that

In

di¡erent

the

degradation

homologous

depolymerization

obtained.

K18

was

system

to

one

pro¢les

x

are

K100/K100

main

depolymerized

oligomer

to

References

several

[1] Geyer, H., Himmelspach, K., Kwiatkowski, B., Schlecht, S.

degradation products. K22 instead was only partially

and Stirm, S. (1983) Degradation of bacterial surface carbo-

cleaved giving rise to a ladder-like pattern. The Hib

hydrates by virus-associated enzymes. Pure Appl. Chem. 55,

polysaccharide was cleaved more e¤ciently than K22 to a series of degradation products.

Two di¡erent

A reasonable background for the phage activity comprises

the

of

polysaccharides

the

chemical

O23 :K18 :H15

and

antigens with (2fering in partial

composition

and

investigated.

O23 :K22 :H15

exhibit

structure

E.

coli

capsular

L-Rib-1.2-Rit-5-P) as backbone, dif-

O-acetylation

at C-3 of the ribose in

the case of K18 [7]. The K100 capsular polysaccharide was reported to be a (3-

L-Rib-1.2-Rit-5-P)

poly-

mer [14], thus di¡ering from K22 by the phosphate linkage to the ribose moiety. The structure of the Hib capsular antigen is known to be a polymer of (3-

L-Rib-1.1-Rit-5-P)

[14]. Thus, the capsular poly-

saccharides investigated here di¡er in some respect in the primary structure. But they have one structural feature ^ ribitol-5-phosphate linkages ^ in common which may be responsible as the primary receptor for

xK100

637^653. [2] Nimmich, W., Schmidt, G. and Krallmann-Wenzel, U. (1991)

and for the speci¢city of the phage-

associated enzymatic hydrolysis of the capsular polysaccharides. Understandably, the ferricyanide method did not give any indication for an increase of the reducing power during the phage-polysaccharide interactions. To explain the di¡erences in the phage

Escherichia coli

polysaccharide depolymerases

each associated with one of the coliphage

xK5

and

xK20.

FEMS Microbiol. Lett. 82, 137^142. [3] Devine, D.A., Robinson, L. and Roberts, A.P. (1989) Occurrence of K1, K5 and O antigens in

Escherichia coli

isolates

from patients with urinary tract infections and bacteraemia. J. Med. Microbiol. 30, 295^299. [4] Nimmich, W. (1994) Detection of

Escherichia coli

K95 strains

by bacteriophages. J. Clin. Microbiol. 32, 2843^2845. [5] Vann, W.F., So ë derstro ë m, T., Egan, W., Tsui, F.-P., Schneerson, R., Òrskov, I. and Òrskov, F. (1983) Serological, chemical, and structural analyses of the

Escherichia coli

cross-reac-

tive capsular polysaccharides. Infect. Immun. 39, 623^629. [6] Òrskov, F. and Òrskov, I. (1984) Serotyping of

coli.

Escherichia

Methods Microbiol. 14, 43^112.

[7] Rodriguez, M.-L., Jann, B. and Jann, K. (1988) Comparative structural elucidation of the K18, K22, and K100 antigens of

Escherichia coli

as related ribosyl-ribitol phosphates. Carbo-

hydr. Res. 173, 243^253. [8] Schneerson, R., Bradshaw, M., Whisnant, J.K., Myerowitz,

Escherichia coli antigen cross-reactive with the capsular polysaccharide of Haemophilus in£uenzae type b : occurrence among known serotypes, and immunochemical and biological properties of E. coli antisera toward H. in£uenzae type B. J. Immunol. 108, R.L., Park, J.C. Jr. and Robbins, J.B. (1972) An

1551^1562. [9] Nimmich,

FEMSLE 7670 20-10-97

W.,

Krallmann-Wenzel,

U.,

Mu ë ller,

B.

and

W. Nimmich / FEMS Microbiology Letters 153 (1997) 105^110

110

Schmidt, G. (1992) Isolation and characterization of bacter-

of 3-deoxy-D-manno-2-octulosonic acid (KDO). Biochem. Bi-

iophages speci¢c for capsular antigens K3, K7, K12, and K13

ophys. Res. Commun. 136, 329^335.

of

Escherichia coli.

Zbl. Bakteriol. 276, 213^220.

[13] Nimmich, W., Voigt, W. and Seltmann, G. (1997) Character-

[10] Imoto, T. and Yagishita, K. (1971) A simple activity measurement of lysozyme. Agr. Biol. Chem. 35, 1154^1156.

ization of urinary

Escherichia coli

O75 strains. J. Clin. Micro-

biol. (in press).

[11] Altmann, F., Ma ë rz, L., Stirm, S. and Unger F.M. (1987) Two

[14] Tsui, F.-P., Egan, W., Summers, M.F., Byrd, R.A., Schneer-

additional bacteriophage-associated glycan hydrolases cleav-

son, R. and Robbins, J.B. (1988) Determination of the struc-

ing ketosidic bonds of 3-deoxy-D-manno-2-octulosonic acid

ture of the

in capsular polysaccharides of

Escherichia coli.

FEBS Lett.

£uenzae.

221, 145^149. [12] Altmann, F., Kwiatkowski, B. and Stirm, S. (1986) A bacteriophage-associated glycanase cleaving

L-pyranosidic

linkages

Escherichia coli

K100 capsular polysaccharide,

cross-reacting with the capsule from type b

Haemophilus in-

Carbohydr. Res. 173, 65^74.

[15] Jennings, H.J. (1990) Capsular polysaccharide vaccines. Curr. Top. Microbiol. Immunol. 150, 97^127.

FEMSLE 7670 20-10-97