Zbl. Bakt. 277, 419-428 (1992 ) © Gustav Fischer Verlag, Srurrgart/New York
Characterization of a Lipopolysaccharide Encoded by a Recombinant Shigella sonnei Plasmid in Escherichia coli K-12 GUNTRAM SEL TMANN I , YURIY A . KNIREL 2, ALEXANDER S . SHA SHK O V 2 , and HELMUT TS CHApE 1 1
2
Robert -Koch-Institu r des Bundesgesundheitsamts, Bereich Wernigerode, 0-3 700 Wernigerode, Germany N. D. Zelinskij Institute of Organic Chemistry, Russian Academy of Sciences, Moscow, Russia
With 5 Figures· Received February 25, 1992 . Revision received Ma y 11, 1992 . Accepted Jun e 29, 1992
Summary Th e genetic inform ation to synthesize the S-specific region of Shigella sonnei phase I lipopol ysaccharide (LPS) is localized on a 180 kb plasmid which is lost quite readily. A recombinant plasmid derivat ive remaining stable in the bacteri a was shown to determine the S-specific region of the LPS which is completely identical with that of a S. sonnei pha se I strai n following tran sfer in Escherichia coli K-12. However, the length control in polysaccharide biosynthesis is lost at least part ially.
Zusammenfassung Shigella sonnei Phase l-Stamme besitzen ein 180 kb-Plasmid , das fur die Biosynthese der S-spezifischen Region des Lipopolysaccharides (LPS) kodiert. Das Plasmid und damit die Phase I-Eigenschaften gehen im allgemeinen schnell verloren. Ein rekombinantes Plasmidderivat bleibt stabil in den Bakterien. In Escherichia coli K-12 bildet es die serologischen S. sonn ei Phase I-Eigenschaften aus. In dieser Arbeit wird gezeigt, daf das Plasmidderivat die Biosynthese einer S-spezifischen LPS-Region determiniert, die mit der eines S. sonnei Phase l-Srammes vollstandig identisch ist. Allerdings ist die Kontrolle der Lange (= Zahl der "r epeating units" ) bei der Biosynthe se der S-spezifischen Region zumindest teilweise verloren gegangen.
Shigella so nne i strai ns generally occur in two modification s named ph ase I and phase II. Phase I represents an S form and is viru lent, phase II repre sents an R form (rype R1 ; 14 ) and is avirulent. Phase I stra ins rather frequently undergo a spontane ou s con ver-
420
G. Seltrnann, Y. A. Knirel, A. S.Shashkov, and H. Tschape
sian to phase II, sometimes one or two passages on nutrient agar surfaces are sufficient to fully produce this effect. This observation is explained by the fact that the genetic determinants encoding for the complete phase I 0 antigen is localized on a 180 kb plasmid (Vir plasmid), which can be lost quite readily (10). The reason for such an instability is unknown. Since similar Vir plasmid segregations have been observed in other bacterial hosts, too (10), either the plasmid replication is insufficient or the plasmid load (e. g. by 0 antigen formation) might be disadvantageous during subcultivation. Observations that phase II segregants often contain deleted Vir plasmid DNA (which is now stably maintained) and that the piece of deleted DNA comprises the genes for invasion (ipa) and for 0 antigen synthesis (ssa; Tscbdpe, unpublished data) would imply that phase I cells of S. sonnei could be overgrown by the rough phase II segregants. However, O-antigen-determining plasmid derivatives have been obtained after cointegrate formation with pTHI0 (18) which are now stably maintained. This acquired stability could be either of plasmid origin or due to an altered 0 antigen structure. In order to decide between both possibilities, we comparatively characterized the 0 antigenic lipopolysaccharides (LPS) synthesized i. by an Escherichia coli K-12 substrain carrying the stable plasmid variant, ii. by the same strain without plasmid, and iii. by the parental S. sonnei strain by several serological and (physico-) chemical methods. Both S-type LPS (i. and iii.) have been shown to be identical in their S-specific but not in their R-specific part. However, the length control in the course of the Sspecific polysaccharide biosynthesis is lost at least partially.
Materials and Methods Strains. E. coli K-12 substrains, S. sonnei wild type strain 9773/63, and the plasmids used throughout these experiments are listed in Table 1. Plasmid transfer procedures. The transfer of S. sonnei Vir plasmid was carried out on solid media at 40°C in the presence of 100 I-lg ampicillin and 50 I-lg kanamycin as described by Watanabe and Nakamura (18), using pTH10 as mobilizing agent. This procedure implies a cointegrate formation via Tn1 which in turn can lead to various recombinant molecules. Recombinant plasmids have been detected in E. coli K-12 strains by agarose gel electrophoresis of plasmid DNA extracted from the respective E. coli K-12 strains encoding for Shigella 0 antigens. These recombinant plasmids were further characterized by standard plasmid methods as described by Tschape et al. (16) and by plasmid DNA fingerprints according to Grinsted and Bennett (5) using the restriction enzymes HindIII, EcoRI, and EcoRV. Furthermore, the recombinant plasmids have been characterized with regard to the presence of the ipa-gene cluster (invasion plasmid antigen) by hybridization studies using the respective gene probe. The plasmid pIE1204 was applied for generation of the ipaB probe (Prager and Tschdpe, unpublished). Hybridization was carried out as described by Grinstead and Bennett (5), under stringent conditions. Origin of sera. All sera (5. sonnei 9773/63 and E. coli IE 1351) were produced according to standard methods (13). E. coli K-12 serum was a kind gift from Lore Brade, Borstel. Isolation of lipopolysaccharides. The bacteria were cultivated in nutrient broth for 6 h at 37°C under intensive stirring and aeration, collected by centrifugation, washed once with saline, and freeze-dried. They were extracted with phenol/water as described by Westphal and ]ann (19). The pooled water phases were dialyzed intensively, then concentrated in vacuo to about 50% and centrifuged 2 hat 105000 g. The sediment was dissolved in water, checked for the absence of materials absorbing at 260 nm, recentrifuged, if necessary and freeze-dried.
LPS Encoded by a Recombinant Plasmid
421
Isolation and purification of a -specific polysaccharides. A 1% solutio n of the LPS in 2% acetic acid was heated for 2 hour s at 100 °C (water bath ). The precipitate was removed by centrifugation. The supernatant was concentrated in vacuo (below 40 °C) and separated by chromatography on a column (46 X 1.7 ern) of Sephadex G-50 in pyridine : acetic acid : water = 4 : 10 : 986 at a flow rate of 1 mllmin. Fractions were collected and analyzed by the orcinol-sulphuric acid reaction with the help of a Technicon Autoan alyzer II sugar detection system at 490 nm. Fract ions conta ining the high-molecular weight material (peak I) were combined, concentrated in vacuo, and freeze-dried. The material s of peak I were dissolved in water and oxidized by 0.1 M sodium metaper iodate (36 h, 20 °C, in the dark ), the excess of oxidizer was destroyed by ethylene glycol, reduced with sodium boro hydride (1 h, 25 °C), dialyzed against distilled wat er and evaporated in vacuo. The residue was heated in 2% acetic acid (2 h, 100 °C), dialyzed against distilled water , and freeze-dried. Immunodiffusion. It was perform ed in agar gels according to Behm (63). 1% solutions of each LP5 (prepared as described by 0rskov and 0rskov, 12) were dropp ed into the outer wells according to the scheme presented in Fig. 1. The centra l wells contained sera against S. sonnei phase I, E. coli K-12, or E. coli IE 1351. The plates were stored at room temperature in a moist chamber for 24 hours and afterwards in saline for 48 hour s to remove unspecific pro teins. After drying, the gels were stained in a 0.5 % solution of Amido black lOB in methanol/acetic acid (8 : 1) and destained in the same solvent . Discontin uous polyacrylamide gel electrophoresis (PAGE) . The method was applied in two modes i.e., - in the presence of sodium dod ecyl sulphate (50 S-PAGE) according to Kusecek et al. (13), however, using 11% acrylamide and 0.2% methylene bisacrylamide without gradient and - in the presence of sodium deoxycholate (DOC-PAGE): the composition of both the gels and the buffer solutions was the same as in the case of SOS-PAGE, however, SOS was substituted by DOC in a fivefold concentration (1). In both cases, bands were visualized by silver staining (18). 13C-NMR spectroscopy. It was performed on a Bruker AM-300 instrument for the solutions in 0 20 at 30 °C using acetone (b = 31.55 ppm) as the internal standard. Result s Characterization of strain s and plasmids S. sonnei 9773 /63 contained a 180 kb DNA mo lecule only, whic h represente d the S. sonnei-spec ific Vir plasmid. Thi s plasmid was transferred, by means of pTH I0, to the E. coli K-12 derivative CV601. Plasmid-containing strains were isolated that expressed the S. sonnei LPS. Among th ese exco njugant E. coli K-12 strai ns, some plasmi d derivatives could be detected that were 125 kb in size only and remained stable under E. coli K-12 conditions. Such a derivat ive plasmid was designated pIE988 (T able 1). As characterized by hybr idization pattern and genetic tests (data not shown), pIE9 88 consists of complete pTHI0 DNA, additionally to a piece of 55 kb Vir plasmid DNA co mprising the det erminant s for LPS synthesis and invasion (ipa). Characterization of LPS by im m unodiffusion On diffusion agai nst a S. sonnet ph ase l-serum, the LPS of both S. sonnei 9773/63 and E. coli IE1351 yielded one single pre cipitation line each, showing complete fusion (Fig. 1). The LPS of E. coli C V60 1 d id not react with this serum. On the oth er han d, this LPS formed a strong line o n diffu sion against the K-12 seru m. Th e same line was formed, with much less intensity by the LPS from E. coli IE1351 , but not by the S. sonnei 977 3/63 LPS. Against seru m lE1351, the LPSs beha ved as against serum 9773/ 63. However, LPS IE 135 1 formed a second (weaker) band , pres um abl y due to the presence of some unsubstituted K-12 LPS and K-12 antibodies.
422
G. Seltmann, Y. A. Knire!, A. S.Shashkov, and H. Tschape
Table 1. Strains and plasmids used throughout this study Designation
Relevant properties
References
Plasmids pTHI0 pIE988
57 Kbp IncPl 125 Kbp IncPl
18 this paper
Tc Am Km Tc Am Km ipa ssa
Strains E.coliK-12 CV601 J53 IE1351
plasmid-free; thr leu thi lac rif" plasmid-free pIE 988 in CV601
16 4 this paper
S.sonnei 9773/63
Vir; drug-sensitive
15
Abbreviations: Tc - tetracycline resistance; Am - ampicillin resistance; Km - kanamycin resistance; rifR - rifampicin resistance; Kbp - Kilobase pairs; ipa - invasion plasmid antigen; ssa - S. sonnei antigen (0 antigen); leu -leucine; thr - threonine; thi - thiamine; laclactose; vir - virulence.
PAGE of the LPS
In SDS-PAGE, both S-rype LPS yielded the typical ladder-like patterns. The arrangement of the bands in the case of both S. sonnei 9773/63 and E. coli IE1351 was identical, indicating the same molecular weight of the repeating units (Fig. 2). However, in the case of strain 9773/63, a bimodal structure was much more pronounced than in the case of strain IE 1351. In DOC-PAGE (Fig. 3), it was clearly visible that the
0\2 o
0(0
~0
o o
o
(0
@(0
,o 0
o
~0
~
CD
CD
CD
Fig. 1. Immunodiffusion of LPS from S. sonnei 9773/63 (1), E. coli IE1351 (2), and E. coli CV601 (3) towards sera against S. sonnei 9773/63 (I; absorbed with phase II bacteria), E. coli K-12 (II), and E. coli IE 1351 (III).
LPS Encoded by a Recombinant Plasmid
423
lipid A-core regions (see 12) of S. sonnei 9773/63 on the one hand and E. coli IE1351 on the other were quite different. On the contrary, the lipid A-core region of the E. coli 1E1351 LPS was very similar to that of the parental strain CV601.
1
2
Fig. 2. SDS-PAGE of LPS from S. sonnei 9773/63 (lane 1) and F. coli IE1351 (lane 2).
123 Fig. 3. DOC-PAGE of LPS from E. coli lE1351 (lane 1), E. coli CV610 (lane 2) and S. sonnei 9773/63 (lane 3).
Gel chromatography of the O-specific polysaccharides on Sephadex G-50 Only the polysaccharides of strains S. sonnei 9773/63 and E. coli IE13S1 were separated. The elution profiles of both polysaccharides were nearly identical (Fig. 4). In both cases, peak I contained the O-specific polysaccharide, however, as could be shown by 13C-NMR analysis, they were contaminated by an a-1.4-glucan (carbon signals at 62.0, 72.7, 72.9, 74.6, 78.9, and 101.0 ppm). Peak II was formed by the Rspecific polysaccharides, and the small peak III, by KDO phosphate. The magnitudes of peaks I and II were comparable in both cases; peak III was more significant in the case of the plasmid-containing E. coli K-12 derivative. Purification of the O-specific polysaccharide The polysaccharides present in peaks I were further purified by periodate oxidation and subsequent mild acid hydrolysis as described in "Materials and Methods". The 0-
424
G. Seltmann, Y. A. Knirel, A. S.Shashkov, and H. Tschape
10 ",1
:
I
Fig. 4. Ge! chromatography on Sephadex G-50 of the O-specific polysaccharides from strains S. sonnei 9773/63 and E. coli IE 1351 (= 1250). For experimental details see "Materials and Methods".
specific polysaccharide proved to be stable under these conditions. After this purification, the sample was free of the glucan (13C-NMR control). 13C_NMR spectroscopic investigations
In Fig. 5, the 13C-NMR spectra in the range between 13 and 111 ppm and 172 and 180 ppm are shown and it is seen that they are identical. Moreover, they are identical with the spectra presented by Kenne er al. (8), by means of which they elucidated the structure of the S-specific region of the S. sonnei LPS. As shown in Table 2, the chemical shifts were in agreement with the proposed chemical structure (see Discussion).
'Uf,\
100
90
80
70
60 PPM
50
40
30
20
Fig, 5 . 13C-NM R spectra of the O-specific polysacchar ides from strai ns S. sonnei 9773/63 (lower lane) and E. coli IE l35l (up per lane).
PPfl
IJiaJI1II
~
V,
tv
s;
[3
~
'""
~
'"
sr S'
3
::0
'" ",.,o
"
Characterization of a Lipopolysaccharide Encoded by a Recombinant Shigella sonnei Plasmid in Escherichia coli K-12 GUNTRAM SEL TMANN I , YURIY A . KNIREL 2, ALEXANDER S . SHA SHK O V 2 , and HELMUT TS CHApE 1 1
2
Robert -Koch-Institu r des Bundesgesundheitsamts, Bereich Wernigerode, 0-3 700 Wernigerode, Germany N. D. Zelinskij Institute of Organic Chemistry, Russian Academy of Sciences, Moscow, Russia
With 5 Figures· Received February 25, 1992 . Revision received Ma y 11, 1992 . Accepted Jun e 29, 1992
Summary Th e genetic inform ation to synthesize the S-specific region of Shigella sonnei phase I lipopol ysaccharide (LPS) is localized on a 180 kb plasmid which is lost quite readily. A recombinant plasmid derivat ive remaining stable in the bacteri a was shown to determine the S-specific region of the LPS which is completely identical with that of a S. sonnei pha se I strai n following tran sfer in Escherichia coli K-12. However, the length control in polysaccharide biosynthesis is lost at least part ially.
Zusammenfassung Shigella sonnei Phase l-Stamme besitzen ein 180 kb-Plasmid , das fur die Biosynthese der S-spezifischen Region des Lipopolysaccharides (LPS) kodiert. Das Plasmid und damit die Phase I-Eigenschaften gehen im allgemeinen schnell verloren. Ein rekombinantes Plasmidderivat bleibt stabil in den Bakterien. In Escherichia coli K-12 bildet es die serologischen S. sonn ei Phase I-Eigenschaften aus. In dieser Arbeit wird gezeigt, daf das Plasmidderivat die Biosynthese einer S-spezifischen LPS-Region determiniert, die mit der eines S. sonnei Phase l-Srammes vollstandig identisch ist. Allerdings ist die Kontrolle der Lange (= Zahl der "r epeating units" ) bei der Biosynthe se der S-spezifischen Region zumindest teilweise verloren gegangen.
Shigella so nne i strai ns generally occur in two modification s named ph ase I and phase II. Phase I represents an S form and is viru lent, phase II repre sents an R form (rype R1 ; 14 ) and is avirulent. Phase I stra ins rather frequently undergo a spontane ou s con ver-
420
G. Seltrnann, Y. A. Knirel, A. S.Shashkov, and H. Tschape
sian to phase II, sometimes one or two passages on nutrient agar surfaces are sufficient to fully produce this effect. This observation is explained by the fact that the genetic determinants encoding for the complete phase I 0 antigen is localized on a 180 kb plasmid (Vir plasmid), which can be lost quite readily (10). The reason for such an instability is unknown. Since similar Vir plasmid segregations have been observed in other bacterial hosts, too (10), either the plasmid replication is insufficient or the plasmid load (e. g. by 0 antigen formation) might be disadvantageous during subcultivation. Observations that phase II segregants often contain deleted Vir plasmid DNA (which is now stably maintained) and that the piece of deleted DNA comprises the genes for invasion (ipa) and for 0 antigen synthesis (ssa; Tscbdpe, unpublished data) would imply that phase I cells of S. sonnei could be overgrown by the rough phase II segregants. However, O-antigen-determining plasmid derivatives have been obtained after cointegrate formation with pTHI0 (18) which are now stably maintained. This acquired stability could be either of plasmid origin or due to an altered 0 antigen structure. In order to decide between both possibilities, we comparatively characterized the 0 antigenic lipopolysaccharides (LPS) synthesized i. by an Escherichia coli K-12 substrain carrying the stable plasmid variant, ii. by the same strain without plasmid, and iii. by the parental S. sonnei strain by several serological and (physico-) chemical methods. Both S-type LPS (i. and iii.) have been shown to be identical in their S-specific but not in their R-specific part. However, the length control in the course of the Sspecific polysaccharide biosynthesis is lost at least partially.
Materials and Methods Strains. E. coli K-12 substrains, S. sonnei wild type strain 9773/63, and the plasmids used throughout these experiments are listed in Table 1. Plasmid transfer procedures. The transfer of S. sonnei Vir plasmid was carried out on solid media at 40°C in the presence of 100 I-lg ampicillin and 50 I-lg kanamycin as described by Watanabe and Nakamura (18), using pTH10 as mobilizing agent. This procedure implies a cointegrate formation via Tn1 which in turn can lead to various recombinant molecules. Recombinant plasmids have been detected in E. coli K-12 strains by agarose gel electrophoresis of plasmid DNA extracted from the respective E. coli K-12 strains encoding for Shigella 0 antigens. These recombinant plasmids were further characterized by standard plasmid methods as described by Tschape et al. (16) and by plasmid DNA fingerprints according to Grinsted and Bennett (5) using the restriction enzymes HindIII, EcoRI, and EcoRV. Furthermore, the recombinant plasmids have been characterized with regard to the presence of the ipa-gene cluster (invasion plasmid antigen) by hybridization studies using the respective gene probe. The plasmid pIE1204 was applied for generation of the ipaB probe (Prager and Tschdpe, unpublished). Hybridization was carried out as described by Grinstead and Bennett (5), under stringent conditions. Origin of sera. All sera (5. sonnei 9773/63 and E. coli IE 1351) were produced according to standard methods (13). E. coli K-12 serum was a kind gift from Lore Brade, Borstel. Isolation of lipopolysaccharides. The bacteria were cultivated in nutrient broth for 6 h at 37°C under intensive stirring and aeration, collected by centrifugation, washed once with saline, and freeze-dried. They were extracted with phenol/water as described by Westphal and ]ann (19). The pooled water phases were dialyzed intensively, then concentrated in vacuo to about 50% and centrifuged 2 hat 105000 g. The sediment was dissolved in water, checked for the absence of materials absorbing at 260 nm, recentrifuged, if necessary and freeze-dried.
LPS Encoded by a Recombinant Plasmid
421
Isolation and purification of a -specific polysaccharides. A 1% solutio n of the LPS in 2% acetic acid was heated for 2 hour s at 100 °C (water bath ). The precipitate was removed by centrifugation. The supernatant was concentrated in vacuo (below 40 °C) and separated by chromatography on a column (46 X 1.7 ern) of Sephadex G-50 in pyridine : acetic acid : water = 4 : 10 : 986 at a flow rate of 1 mllmin. Fractions were collected and analyzed by the orcinol-sulphuric acid reaction with the help of a Technicon Autoan alyzer II sugar detection system at 490 nm. Fract ions conta ining the high-molecular weight material (peak I) were combined, concentrated in vacuo, and freeze-dried. The material s of peak I were dissolved in water and oxidized by 0.1 M sodium metaper iodate (36 h, 20 °C, in the dark ), the excess of oxidizer was destroyed by ethylene glycol, reduced with sodium boro hydride (1 h, 25 °C), dialyzed against distilled wat er and evaporated in vacuo. The residue was heated in 2% acetic acid (2 h, 100 °C), dialyzed against distilled water , and freeze-dried. Immunodiffusion. It was perform ed in agar gels according to Behm (63). 1% solutions of each LP5 (prepared as described by 0rskov and 0rskov, 12) were dropp ed into the outer wells according to the scheme presented in Fig. 1. The centra l wells contained sera against S. sonnei phase I, E. coli K-12, or E. coli IE 1351. The plates were stored at room temperature in a moist chamber for 24 hours and afterwards in saline for 48 hour s to remove unspecific pro teins. After drying, the gels were stained in a 0.5 % solution of Amido black lOB in methanol/acetic acid (8 : 1) and destained in the same solvent . Discontin uous polyacrylamide gel electrophoresis (PAGE) . The method was applied in two modes i.e., - in the presence of sodium dod ecyl sulphate (50 S-PAGE) according to Kusecek et al. (13), however, using 11% acrylamide and 0.2% methylene bisacrylamide without gradient and - in the presence of sodium deoxycholate (DOC-PAGE): the composition of both the gels and the buffer solutions was the same as in the case of SOS-PAGE, however, SOS was substituted by DOC in a fivefold concentration (1). In both cases, bands were visualized by silver staining (18). 13C-NMR spectroscopy. It was performed on a Bruker AM-300 instrument for the solutions in 0 20 at 30 °C using acetone (b = 31.55 ppm) as the internal standard. Result s Characterization of strain s and plasmids S. sonnei 9773 /63 contained a 180 kb DNA mo lecule only, whic h represente d the S. sonnei-spec ific Vir plasmid. Thi s plasmid was transferred, by means of pTH I0, to the E. coli K-12 derivative CV601. Plasmid-containing strains were isolated that expressed the S. sonnei LPS. Among th ese exco njugant E. coli K-12 strai ns, some plasmi d derivatives could be detected that were 125 kb in size only and remained stable under E. coli K-12 conditions. Such a derivat ive plasmid was designated pIE988 (T able 1). As characterized by hybr idization pattern and genetic tests (data not shown), pIE9 88 consists of complete pTHI0 DNA, additionally to a piece of 55 kb Vir plasmid DNA co mprising the det erminant s for LPS synthesis and invasion (ipa). Characterization of LPS by im m unodiffusion On diffusion agai nst a S. sonnet ph ase l-serum, the LPS of both S. sonnei 9773/63 and E. coli IE1351 yielded one single pre cipitation line each, showing complete fusion (Fig. 1). The LPS of E. coli C V60 1 d id not react with this serum. On the oth er han d, this LPS formed a strong line o n diffu sion against the K-12 seru m. Th e same line was formed, with much less intensity by the LPS from E. coli IE1351 , but not by the S. sonnei 977 3/63 LPS. Against seru m lE1351, the LPSs beha ved as against serum 9773/ 63. However, LPS IE 135 1 formed a second (weaker) band , pres um abl y due to the presence of some unsubstituted K-12 LPS and K-12 antibodies.
422
G. Seltmann, Y. A. Knire!, A. S.Shashkov, and H. Tschape
Table 1. Strains and plasmids used throughout this study Designation
Relevant properties
References
Plasmids pTHI0 pIE988
57 Kbp IncPl 125 Kbp IncPl
18 this paper
Tc Am Km Tc Am Km ipa ssa
Strains E.coliK-12 CV601 J53 IE1351
plasmid-free; thr leu thi lac rif" plasmid-free pIE 988 in CV601
16 4 this paper
S.sonnei 9773/63
Vir; drug-sensitive
15
Abbreviations: Tc - tetracycline resistance; Am - ampicillin resistance; Km - kanamycin resistance; rifR - rifampicin resistance; Kbp - Kilobase pairs; ipa - invasion plasmid antigen; ssa - S. sonnei antigen (0 antigen); leu -leucine; thr - threonine; thi - thiamine; laclactose; vir - virulence.
PAGE of the LPS
In SDS-PAGE, both S-rype LPS yielded the typical ladder-like patterns. The arrangement of the bands in the case of both S. sonnei 9773/63 and E. coli IE1351 was identical, indicating the same molecular weight of the repeating units (Fig. 2). However, in the case of strain 9773/63, a bimodal structure was much more pronounced than in the case of strain IE 1351. In DOC-PAGE (Fig. 3), it was clearly visible that the
0\2 o
0(0
~0
o o
o
(0
@(0
,o 0
o
~0
~
CD
CD
CD
Fig. 1. Immunodiffusion of LPS from S. sonnei 9773/63 (1), E. coli IE1351 (2), and E. coli CV601 (3) towards sera against S. sonnei 9773/63 (I; absorbed with phase II bacteria), E. coli K-12 (II), and E. coli IE 1351 (III).
LPS Encoded by a Recombinant Plasmid
423
lipid A-core regions (see 12) of S. sonnei 9773/63 on the one hand and E. coli IE1351 on the other were quite different. On the contrary, the lipid A-core region of the E. coli 1E1351 LPS was very similar to that of the parental strain CV601.
1
2
Fig. 2. SDS-PAGE of LPS from S. sonnei 9773/63 (lane 1) and F. coli IE1351 (lane 2).
123 Fig. 3. DOC-PAGE of LPS from E. coli lE1351 (lane 1), E. coli CV610 (lane 2) and S. sonnei 9773/63 (lane 3).
Gel chromatography of the O-specific polysaccharides on Sephadex G-50 Only the polysaccharides of strains S. sonnei 9773/63 and E. coli IE13S1 were separated. The elution profiles of both polysaccharides were nearly identical (Fig. 4). In both cases, peak I contained the O-specific polysaccharide, however, as could be shown by 13C-NMR analysis, they were contaminated by an a-1.4-glucan (carbon signals at 62.0, 72.7, 72.9, 74.6, 78.9, and 101.0 ppm). Peak II was formed by the Rspecific polysaccharides, and the small peak III, by KDO phosphate. The magnitudes of peaks I and II were comparable in both cases; peak III was more significant in the case of the plasmid-containing E. coli K-12 derivative. Purification of the O-specific polysaccharide The polysaccharides present in peaks I were further purified by periodate oxidation and subsequent mild acid hydrolysis as described in "Materials and Methods". The 0-
424
G. Seltmann, Y. A. Knirel, A. S.Shashkov, and H. Tschape
10 ",1
:
I
Fig. 4. Ge! chromatography on Sephadex G-50 of the O-specific polysaccharides from strains S. sonnei 9773/63 and E. coli IE 1351 (= 1250). For experimental details see "Materials and Methods".
specific polysaccharide proved to be stable under these conditions. After this purification, the sample was free of the glucan (13C-NMR control). 13C_NMR spectroscopic investigations
In Fig. 5, the 13C-NMR spectra in the range between 13 and 111 ppm and 172 and 180 ppm are shown and it is seen that they are identical. Moreover, they are identical with the spectra presented by Kenne er al. (8), by means of which they elucidated the structure of the S-specific region of the S. sonnei LPS. As shown in Table 2, the chemical shifts were in agreement with the proposed chemical structure (see Discussion).
'Uf,\
100
90
80
70
60 PPM
50
40
30
20
Fig, 5 . 13C-NM R spectra of the O-specific polysacchar ides from strai ns S. sonnei 9773/63 (lower lane) and E. coli IE l35l (up per lane).
PPfl
IJiaJI1II
~
V,
tv
s;
[3
~
'""
~
'"
sr S'
3
::0
'" ",.,o
"
