0019-9567/78/0019-0231$02.00/0
Vol. 19, No. 1
INFECTION AND IMMUNITY, Jan. 1978, p. 231-238
Printed in U.S.A.
Copyright i) 1978 American Society for Microbiology
Phage-Induced Changes in Vibrio cholerae:. Serotype and Biotype Conversions JAMES E. OGG,l* MADAN B. SHRESTHA,2 AND LAXMAN POUDAYL2 Department of Microbiology, Colorado State University, Fort Collins, Colorado 80523,1 and the Central Health Laboratory, Kathmandu, Nepal' Received for publication 25 May 1977
Phage infection of Vibrio cholerae resulted in antigenic changes. A strain of biotype cholerae serotype Ogawa was converted into serotype Hikojima and gained the ability to synthesize antigenic factor C. Some phage-converted strains remained stable after subculture and were immune to superinfection with the same phage. The stable converted strains were lysogenic and released phage having a host range similar to the phage of the donor strain. Reinfection of unstable converted strains which had "lost" antigen C yielded types able to again synthesize this antigen. The polymyxin resistance character was expressed in the biotype cholerae strain after infection with some phage preparations. These polymyxin-resistant strains possessed three main characteristics of El Tor vibrios. The phage-induced changes described provide V. cholerae with the potential for innumerable genetic combinations.
Lysogeny in Vibrio cholerae biotype eltor was first reported by Newman (Bacteriol. Proc., p. 77, 1960) and later by others (4, 18). The general extent of lysogeny in biotype eltor and the morphological and biological properties of El Tor phages have been investigated (7, 19, 21). A correlation exists between pathogenicity and lysogeny in both the El Tor and classical strains of V. cholerae (9, 18). Newman and Eisenstark (7) suggested that phage conversion may account for the hemolytic properties of biotype eltor; they also found that V. cholerae biotype cholerae may be lysogenized with temperate phage from El Tor strains. In this paper we present evidence of bacteriophage conversion affecting the serotype of V. cholerae biotype cholerae. What appeared to be transduction of the polymyxin resistance character by phage from a biotype eltor to a biotype cholerae also occurred with some phage preparations. Stable lysogenization of the host strain was demonstrated by (i) i,munity of the converted strain to superinfection with the same phage after five single colony subcultures in broth and (ii) production of phage by the converted strains after the fifth subculture, which had a host range similar to the phage from the donor strain. MATERIALS AND METHODS Bacterial strains. The V. cholerae strains were isolated at the Central Health Laboratory, Kathmandu, Nepal, from stool specimens collected from clinical cases of Asiatic Cholera. Stock cultures were
maintained on nutrient agar slants held at room temperature. The main characteristics of the strains are listed in Table 1. The biotype determination was made on the basis of sensitivity to 50 IU of polymyxin B (Difco) using the disk test (5), the Voges-Proskauer reaction performed on cultures grown in MR/VP medium (Oxoid) at 22 and 37°C, and the chicken erythrocyte agglutination slide test (3). Strains designated biotype eltor were not sensitive to polymyxin B, were Voges-Proskauer positive at both 22 and 37°C, and agglutinated chicken erythrocytes. Strain 029, classified as biotype cholerae, was sensitive to polymyxin B, gave a weakly positive Voges-Proskauer reaction at 37°C and a negative Voges-Proskauer test at 22°C, and did not agglutinate chicken erythrocytes. V. cholerae polyvalent Ogawa and Inaba agglutinating serums supplied by Wellcome Reagents Limited, Wellcome Research Laboratories, Beckenham, England, were used routinely in slide agglutination tests to determine the serotype of strains and the phageconverted isolants. Bacto-V. cholerae Inaba, Ogawa, and Hikojima antiserums (Difco) served as a second source of typing sera in both slide and tube agglutination tests to double check the serotypes of the strains listed in Table 1, as well as the serotypes of some of the phage-converted isolants. Cells of Hikojima serotype sometimes exhibited a slight delay in agglutination in polyvalent and specific absorbed Inaba and Ogawa antisera. Media. Nutrient broth (Oxoid) and nutrient agar (Oxoid) at pH 7.4 were used for the cultivation of V. cholerae and for bacteriophage propagation. The medium used for carbohydrate fermentation tests was Andrade peptone water (Oxoid) containing 0.5% carbohydrate. Isolation and propagation of bacteriophage. The two biotype eltor serotype Hikojima strains 1633
231
232
OGG, SHRESTHA, AND POUDAYL
INFECT. IMMUN.
TABLE 1. Characteristics of V. cholerae strains Fermentation (without gas) after 48 h of: Strain
Biotype
Serotype
0 Factors Glucose
Mannose
Sucrose
Lactose
029 cholerae + + + + (72 h) Ogawa AB 1621 eltor + Ogawa + + AB + (96 h) 1633a eltor Hikojima + ABC + + + 2001a eltor Hikojima ABC + + + + a Strains designated as serotype Hikojima, as described in Bergey's Manual (14), agglutinate in both Ogawaand Inaba-absorbed antisera.
and 2001 were selected as phage donors from 24 strains screened for evidence of lysogeny using the cross-lysis technique. Phage was released from all but 4 of the 24 strains, including strains 029 and 1621, using the following methods. Supernatants were collected from broth cultures which were variously (i) 24 h old; (ii) irradiated for 4 min with a Sylvania (G15T8) germicidal lamp at a distance of 45 cm, diluted with 2 volumes of sterile broth, and incubated for 7 h at 370C, and (iii) held at 560C for 1 h, diluted with 2 volumes of broth, and incubated for 7 h at 37°C (11). The supernatants were screened by spotting a lawn of bacteria prepared with cells suspended in 0.6% nutrient agar containing 2 x 102 M CaCl2. Strain 1621 was an indicator for phage from 12 of the 20 lysogenic strains. The phage from lysogen 1621 did not lyse cells of the other three strains listed in Table 1. Single turbid plaques formed by phage from either strain 1633 or 2001 on lawns of strain 1621 were picked, and each was placed in 1 ml of nutrient broth. Ten milliliters of a 7- to 8-h broth culture of strain 1621 was inoculated with 0.1 ml of a single plaque phage suspension and incubated overnight at 370C. The cells were sedimented in a centrifuge, and the supernatant was collected. In one experiment, strain 029 was substituted for 1621 for the isolation and propagation of phage from strains 1633 and 2001. The agar overlay method (1) was used to determine phage titer and to obtain phage preparations of high titer. The phage preparations were filtered through a membrane filter (0.45-,m pore diameter). The phage lysates were tested before use to confirm cell-free preparations by plating on nutrient agar and incubating the plates for 48 h. Procedure for phage conversion experiments. Two methods were used to infect recipient bacteria with phage from donor strains 1633 and 2001. The broth method was used in most experiments and consisted of mixing 0.2 ml of cell-free phage suspension (titer, about 2 x 109 to 3 x 109 plaque-forming units per ml) with 0.8 ml of bacterial cells from an overnight culture incubated at 370C (viable cell count, about 1 x 109 to 4 x 109 cells per ml) in a sterile vial containing 1 ml of nutrient broth. Sterile CaCl2 solution was added to give a final concentration of 2 x 10-2 M. Two controls were prepared: one with recipient bacteria in nutrient broth containing CaCl2 to check for possible spontaneous changes in serotype and one containing phage in nutrient broth as a double check on the phage preparation being cell free. The cells in broth control were treated in the same manner as the phage-treated culture. The phage in nutrient broth
control was incubated at 370C for 24 h, and 0.1 ml was spotted onto a nutrient agar plate, which was then incubated for 48 h. The phage-treated cells were incubated at room temperature for 3 and then for 4 h at 370C. A loopful of phage-treated culture was streaked onto an agar plate. After overnight incubation at 370C, isolated colonies were picked at random, and each was placed in 10 ml of broth. After incubation, a loopful of each culture was spread over the surface of an agar plate. A polymyxin B disk was placed in the center of the inoculum, and the plate was incubated overnight. The agar cultures were read for sensitivity or resistance to polymyxin B, and the serotype(s) was determined. A small amount of growth from agar plate cultures showing either a change in serotype or a change in the polymyxin reaction was inoculated into 10 ml of broth and incubated overnight. A nutrient agar plate was streaked with a loopful of broth culture and incubated. Isolated colonies were streaked onto nutrient agar slants to establish stock cultures from single colony isolants. These stock cultures were coded with a number followed by "TP" to indicate treated with phage. A second method (the plate method) was the use of a soft agar overlay containing 0.2 ml of phage and 0.3 ml of an overnight broth culture mixed in 3 ml of agar containing 2 x 10-2 M CaCl2. After incubation of the plates for 2 h at room temperature and then for 24 h at 370C, isolated colonies from areas of confluent lysis were picked, and each was placed in 10 ml of broth. After incubation, this broth culture was tested for its polymyxin reaction and serotype(s) using the same procedure as in the broth method. Stock cultures of the phage-treated strains were established. Procedure for testing stability of phage-converted characteristics. A loopful of growth was transferred from a stock culture coded TP into 10 ml of nutrient broth. After overnight incubation at 370C, the culture was streaked onto nutrient agar to obtain isolated colonies. A single colony was picked and used to establish another broth subculture. Each subculture was screened for retention of the phage-induced char-
acteristic(s). Test for lysogenicity and immunity of the phage-converted strains. Lysogenicity of the TP phage-converted strains was determined after the fifth broth subculture by two methods: immunity to infection with homologous phage and release of phage with a host range similar to that of phage donor strains. Immunity to homologous phage was determined by cross-streaking nutrient broth cultures across a band of phage from a donor strain on nutrient agar as well
PHAGE-INDUCED CHANGES IN V. CHOLERAE
VOL. 19, 1978
as by spotting a lawn of a TP strain with a donor phage preparation. The supernatant from 24-h broth cultures and the supernatant collected from UV-induced TP strains were tested for the release of phage by spotting lawns of "indicator" strains of bacteria.
RESULTS Strain 029 (V. cholerae biotype cholerae serotype Ogawa), when treated with phage preparations obtained from strain 1633 (V. cholerae biotype eltor serotype Hikojima), exhibited changes in serotype and/or polymyxin resistance (Table 2). Strain 029 was treated with phage released from a heat-induced culture of 1633, and 11 cultures were established from single colonies. Of these, eight exhibited changes: two gained only the polymyxin resistance characteristic, five gained both polymyxin resistance and the ability to agglutinate in Inaba typing serum, and one remained polymyxin sensitive but gained the ability to agglutinate in Inaba typing serum. (Agglutination in Inaba typing serum TABLE 2. Changes induced in V. cholerae biotype cholerae (strain 029) by phage obtained from strain 1633 (biotype eltor) Culture codea
Poly- Agglutination reaction myxin rec
tio b
Serotype
0 Factors
Ogawa Ogawa Hikojima
AB AB ABC
Ogawa Inaba
Controls 029 1621 1633 Isolants
S R R
+ + +
-
+
(BH) + R AB ITPa Ogawa + + S Hikojima ABC 3TP + R AB 5TP Ogawa + R + Hikojima ABC 6, 7TP Isolants (PH) + + R 26-28TP Hikojima ABC Isolants (BU) + R AB Ogawa 8, 9TP + + S lOTP Hikojima ABC + + R Hikojima ABC 1lTP + AR R Ogawa 12TP + R + Hikojima ABC 15TP R + AB Ogawa 16TP Isolants (PU) + + S 29TP Hikojima ABC + + R 3OTP Hikojima ABC + R AB Ogawa 31TP + + S Hikojima ABC 32TP + + R 33TP Hikojima ABC aTP indicates cultures established from single colonies after phage treatment. The phage was propagated on strain 1621. Only isolants exhibiting characteristics different from strain 029 are listed. B, broth method; P, plate method; H, heat treatment of 1633 to release phage; U, treatment of 1633 to release phage. b S, Sensitivity to polymyxin; R, resistance to polymyxin.
233
indicates that the cells have the ability to synthesize antigenic factor C.) Similar results were obtained using a phage preparation from an UVinduced strain of 1633. Twelve of fourteen isolants had undergone changes: five gained only the polymyxin resistance characteristic, four were polymyxin resistant and exhibited antigenic factor C production, and three remained polymyxin sensitive but agglutinated in Inaba typing serum. In a later independent experiment using the broth method and phage released from strain 1633 after heat treatment, six out of ten cultures established from single colonies gained the ability to agglutinate in Inaba typing serum, but none exhibited polymyxin resistance. The frequency of occurrence of the polymyxin resistance characteristic was not determined in these experiments, but the results indicate that the polymyxin resistance character appeared to be transferred independently of antigenic factor C. Significant was the observation that antigenic factor C occurred in high frequency in phagetreated cells. This level of expression of a serotype characteristic suggests that antigenic factor C production is associated with the infection of 029 cells with phage from 1633. The phage-induced changes of selected isolants were subjected to stability testing, and the polymyxin resistance characteristic (pol+) was found to be unstable in most but not all of the TP cultures (Table 3). Of the four isolants tested which had gained only the polymyxin resistance character, one (31TP) retained the property through five subcultures. One isolant (26TP) which had the dual characteristic (C pol+) retained its polymyxin resistance character but lost antigenic factor C on the fifth subculture. The polymyxin resistance character may be transduced by the phage from strain 1621 into strain 029. If so, in most cases the exogenote apparently is not integrated into the endogenote and the polymyxin resistance character does not persist in the unstable heterogenote. The two TP strains (26TP, 31TP) which retained the polymyxin resistance character through a number of subcultures became useful later for studies on biotype changes. A phage-induced characteristic was considered stable if it survived five subcultures. Of the eight TP isolants screened exhibiting the dual characteristic (C pol+), six retained antigenic factor C, even though seven had lost the polymyxin resistance character. Two of the four isolants which had been converted only to Hikojima serotype maintained the characteristic. The production of antigen C appeared to be correlated with the presence of the 1633 phage genome in the converted strains. The loss of antigenic factor C in some TP isolants could have
234
OGG, SHRESTHA, AND POUDAYL
INFECT. IMMUN.
TABLE 3. Stability ofphage-induced characteristics in strain 029 Culture codea
Controls: 029 1633 1621 Isolants: 3TP 7TP 8TP 9TP lOTP 11TP 15TP 16TP 26TP 27TP 28TP 29TP
Phage-induced characteristicsb
Characteristics- after subculture no.: 1
2
NC NC NC
NC NC NC
NC NC NC
4
5
NC NC NC
NC NC NC
C C pol+ C C C C C pol+ pol+ C C C C C C C pol+ C C C C C C C C C C pol+ pol+ C pol+ C pol+ C pol+ C pol+ C pol+ pol+ C pol+ C C C C C C pol+ C pol+ C C C C C C C C C C C pol+ C C C C C 3OTP 31TP pol+ pol+ pol+ pol+ pol+ pol+ C C 32TP C pol+ C C C C C 33TP aTP indicates a culture established from a single colony after treatment of strain 029 with phage from strain 1633. b C, Gain of antigenic factor C; pol+, resistance to polymyxin B. e-, Characteristics the same as strain 029; NC, no change in original characteristics.
been due to an unstable condition and eventual loss of the phage genome in those cultures. Those cells no longer lysogenic for the 1633 phage genome should become susceptible to reinfection with this particular phage. This was confirmed by experiments where two of the TP isolants tested which had lost antigenic factor C on transfer (15TP and 26TP) regained the ability to agglutinate in Inaba typing serum after treatment with phage from a heat-induced culture of 1633. Following this second phage treatment, two out of five of the 15TP cultures and six out of six of the 26TP cultures established by single colony isolation agglutinated in Inaba typing serum. One of the 15TP isolants converted to serotype Hikojima was also resistant to polymyxin B. Those TP cultures which maintained the antigenic factor C on transfer and carry the 1633 phage genome should be lysogenic as well as immune to infection with homologous phage. Table 4 shows the results from spotting phage preparations onto lawns of bacteria. These immunity and lysogeny tests indicate that those strains which had antigenic factor C after five subcultures in broth were immune to infection with phage from strain 1633 (the original phage donor) and also to phage from 029 (the original recipient strain). The phage released by the stable "converted strains" had the same host
range as the phage from strain 1633. The original recipient strain 029 was lysed by the phage released from the stable antigenic factor C convertants, but not by the phage from strains 26TP and 32TP which had lost the ability to produce antigen C. These two isolants had become nonlysogenic for the 1633 phage, but still retained the prophage carried by the original 029 strain. The phage released by these strains as well as by strains 31TP and 029 (neither of which possessed antigenic factor C) lysed only indicator strain 1621. Therefore, those strains converted to Hikojima serotype apparently still have the original 029 prophage as well as the 1633 prophage. The polymyxin resistance character alone does not confer immunity to phage 1633. Isolants 26TP and 31TP, in which the polymyxin resistance character was maintained, were lysed by phage from strain 1633 and the stable antigen C factor convertants. Phage from strain 2001, like phage from strain 1633, had the ability to convert strain 029 as well as TP isolants 15, 26, and 31 to antigen C production (Table 5). The phage used was released from strain 2001 after UV light treatment and propagated on strain 1621. The broth method was used in treating the recipient cells with phage. Phage 2001 was very effective in converting all the test strains to Hikojima sero-
PHAGE-INDUCED CHANGES IN V. CHOLERAE
VOL. 19, 1978
235
TABLE 4. Immunity and lysogeny in various V. cholerae strainsa
source)n source) 029 1633 7TP
chokrae strain used for lawn
V.
Phage
27TP
29TP
(pol+)
(C)
28TP
(C)
3OTP
(C)
(pol+)
32TP
1621b
-
-
-
-
-
-
-
-
-
-
-
+
-
-
-
-
+
+
+ +
+
+
+
029b 029b
1633b 1633b
7TP
-
-
+ +
-
(C)
1OTP
26TP
(C)
(C)
-+
31TP
+ + + + + lOTP + _ _ 26TP + + + + + 27TP + + + + + 28TP + + + + + 29TP + + + _ + + 3OTP + 31TP t 32TP a TP indicates a culture established from a single colony after treatment of 029 with phage from strain 1633. Symbol in parentheses indicates the phage-induced character retained after five subcultures: C, antigenic factor C; pol+, resistance to polymyxin B. +, Lysis; -, no lysis. b Control. -
-
-
-
-
-
-
-
-
-
-
-
-
-
-
TABLE 5. Changes induced in V. cholerae biotype cholerae strains 029 and 15TP and in polymyxinresistant derivatives 26TP and 31TP by phage from strain 2001. Culture code
Poly- Agglutination myxin reactiona
Serotype
0 Factors
Ogawa Ogawa
AB AB AB AB
Ogawa Inaba
Controls' 029 26TP 31TP 15TP Isolants from 26TP
S R R S
+ + + +
-
7OTP 71-74TP Isolants from 31TP 75-77TP Isolants from
R R
+ +
+
Ogawa Hikojima
AB ABC
R
+
+
Hikojima
ABC
S S S
+ +
+ -
Hikojima Ogawa
ABC
+
+
Hikojima
ABC
Ogawa
Ogawa
029
8OTP 81TP
82-88TP
-
AB
Isolants from 15TP + + Hikojima ABC 120-124TP S a S, Sensitivity to polymyxin B; R, resistant to polymyxin B. bStrains 15TP, 26TP, and 31TP were isolated originally from strain 029 subsequent to infection with phage from strain 1633.
type. Again, the frequency of the occurrence is typical of a phage conversion system. No polymyxin-resistant isolants were recovered from the phage-treated culture of 029 or strain 15TP, but relatively few cells were screened. Three of the
four phage-converted strains of 26TP, five of the six tested phage-converted strains of 029 and all five of those from 15TP remained stable after five transfers. Only one of the phage-converted strains of 31TP was tested, and it lost antigen C on the third broth subculture. Phages from 1633 and 2001 propagated on strain 029 had the ability to convert strains 029, 15TP, 26TP, and 31TP to Hikojima serotype. Thus, the strain on which the phage was propagated did not apparently affect the expression of the 1633 or 2001 phage genome in a recipient cell. Also, strain 029, subsequent to infection with phage in a supernatant collected from a UV-induced culture of strain 1633, yielded isolants which agglutinated in both Inaba and Ogawa typing serums. These results indicate the 1633 phage genome caused induction of antigenic factor C production by cells of strain 029. As described previously, phage from strain 1633 apparently transduced the polymyxin resistance character from a biotype eltor into biotype cholerae strain 029. The question arose whether other biotype eltor characteristics were associated with a strain being resistant to polymyxin B. The stable polymyxin-resistant "transductants" of strain 029 were tested for ability to agglutinate chicken erythrocytes and ability to produce acetylmethylcarbinol at 22°C (Table 6). The polymyxin-resistant strains (26TP, 31TP) agglutinated chicken erythrocytes and were Voges-Proskauer positive at 220C. These transductants then possessed three main characteristics that distinguish the biotype eltor from biotype cholerae (14). Some of the phageconverted strains which possessed the polymyxin resistance character on original isolation but lost this character on transfer (strains 7TP,
236
OGG, SHRESTHA, AND POUDAYL
TABLE 6. Voges-Proskauer and chicken erythrocyte agglutination reactions by polymyxin-resistant transductants and other strains of V. cholerae Strain
Controls 029
Chicken
Voges-Proskauer reac-
erythro-
tiona at: 22°C 37°C
cyte agglutination
-
Slightly +
Polymyxin reaction
-
S
+ + +
R R R
in 18 h 1633 2001 1621 Isolants 26TPC
+ + +
+
+
31TPC
+
+
+
R R
7TP
-
-
S
+ + +
+
Slightly + in 2 h
-
Slightly + in
-
S
15TP
-
1h Slightly + in 2h
-
S
27TP 28TP 29TP
-
+
-
S
+
-
Slightly + in
-
S S
30TP
-
Slightly + in
-
S
1OTP
2h
2h a +, Positive for acetylmethylcarbinol;-, negative 24 h after adding the reagents. b S, Sensitivity to polymyxin B; R, resistant to polymyxin B. c
Stable polymyxin-resistant transductant of strain
029.
15TP, 27TP, 28TP, and 3OTP) did not exhibit the distinguishing characteristics of biotype eltor. The stable antigen C convertants (10 and 29TP) also did not demonstrate any biotype eltor reactions.
DISCUSSION Phage conversion leading to changes in the serotype of V. cholerae has been demonstrated. Cells of the converted strains (derived from a serotype Ogawa treated with phage from strains of serotype Hikojima) agglutinated in Inaba as well as Ogawa typing sera, indicating a gain in ability to synthesize antigenic factor C, a typespecific factor of Inaba serotype (14). This system of phage conversion influencing the production of an antigenic factor is similar to that reported in Salmonella (20) and in Campylobacter fetus (formerly designated V. fetus) by Ogg and Chang (8). However, in C. fetus and in some of the phage-converted Salmonella, one of the cell's original antigenic factors was lost or repressed on subculture. The stable converted V. cholerae strains maintained their original type-specific antigens as well as the "new" antigenic factor C after five subcultures in broth.
INFECT. IMMUN.
The data indicate that the loss of antigenic factor C in the unstable phage-converted isolants was associated with a change from the lysogenic to the nonlysogenic state in the cells for the original converting phage (Table 4, strains 26TP, 32TP). The Ogawa-to-Hikojima serotype change was induced in biotype cholerae by phage released from strains 1633 and 2001. Strains 1633 and 2001, even though of the same serotype, were isolated from clinical cases of cholera at different times. The temperate phage released by these two strains may be the same or different, but they have the same effect on the host cell: the induction of antigen C production. This system of phage conversion may help explain some reports of serotype conversion in V. cholerae occurring under various conditions. Sakazaki and Tamura (13) suggested that serotype Inaba originated from Ogawa by loss of somatic antigenic factor B; they also recovered Hikojima variants from Ogawa cultures grown in homologous specific factor serum. Conversions from Ogawa to Inaba and from Inaba to Ogawa were observed by Sack and Miller (12) in mono-contaminated gnotobiotic mice. They suggested that such serotype changes were due to the selective effect of antibody within the lumen of the intestine. Earlier, Shrivastava and White (16) isolated Inaba and rough strains after cultivation of serotype Ogawa in monospecific antiserum, whereas predominantly Inaba-like serotypes (which possessed an imperfect Ogawa factor) yielded Ogawa types in Inaba antiserum. Serotype conversions in cholera patients were reported by Gangarosa et al. (6). Loss or gain of antigenic factors in V. cholerae may be due to the expression of a phage genome(s), the loss of a prophage and the selection of nonlysogenic cells as could occur under various culture conditions and in the presence of type-specific antibody, or the infection of a cell by phage released by a strain having the same or a different serotype. As shown by the results of our experiments, cells which become nonlysogenic for a particular temperate phage may regain competency to synthesize a specific antigenic factor (Table 5, strains 15TP, 26TP). When exposed to type-specific antibody, one may expect that the same type of change from lysogenic to nonlysogenic would occur in some cells of V. cholerae as obtained in Salmonella (20) and C. fetus
(8).
More than one type of prophage may become part of the genetic complement of V. cholerae (Table 4). Strain 029, which is lysogenic for a phage which lyses 1621, also became lysogenic for another type of phage when infected with phage from strain 1633. How many prophages
VOL. 19, 1978
PHAGE-INDUCED CHANGES IN V. CHOLERAE
may become part of the genetic complement of this organism is unknown. But the possible interactions that may occur between phage genes and between prophage and host chromosome genes confer onto V. cholerae the potential for numerous genetic and phenotypic variations. Phage conversion could provide additional information about phage genome effects on the expression of host chromosomal genes in this important bacterial pathogen. The polymyxin resistance character also was exhibited in V. cholerae biotype cholerae subsequent to infection by temperate phage from V. cholerae biotype eltor. Associated with the polymyxin resistance character were two other characteristics typical of El Tor biotype: the ability to agglutinate chicken erythrocytes and production of acetylmethylcarbinol at 220C. Although the polymyxin resistance proved to be highly unstable in the transductants, two did retain the characteristic. Until the last cholera pandemic, El Tor vibrios were usually associated with localized outbreaks of cholera-like disease and were not considered epidemic strains like classical cholera vibrios (2). The pandemic strain of V. cholerae biotype eltor might have originated from a classical cholera vibrio that was transduced to polymyxin resistance with phage from a biotype eltor or it could have resulted from a mutation to polymyxin resistance in a classical vibrio. The predominant serotype of V. cholerae may change from one epidemic to another (10). Shrivastava (15) noted changes in dominant serotypes in Calcutta (Ogawa in 1964 versus Inaba in 1962 and 1963) and postulated that such changes could be due to conversion of the antigenic patterns of vibrio strains in vivo or to possible conversions under in vitro environmental conditions. Local conversions in the strains of vibrio rather than importation from outside could account for strains of different characteristics. In Nepal, fluctuations in the dominant serotype also have been occurring from one period to another (unpublished observations of the authors). Phage conversion as demonstrated in our work could be one source of such antigenic changes as occur in nature, since it is a highly effective system for introducing genetic information into a cell of V. cholerae. It may account for the origin of highly virulent strains which cause explosive and severe epidemics threatening the health of over half of the world's population. Smith and Goodner stated in 1965 (17), "... .we are forced to the supposition that a vibrio may well contain latent information toward several syntheses as regard antigens and enzymes, but we have no understanding of the force which
237
may activate these bits of stored information. One must also consider that in nature the vibrios probably have access to a great amount of information from other sources." We feel that the data presented in our paper identify one such source of genetic information available to the vibrios: the influence of temperate phage genomes on the synthesis of type-specific antigens and their ability to serve as a passive carrier of host genes. It also is well known that the introduction of virus genetic material into a cell can activate the expression of latent genetic information of the host chromosome. ACKNOWLEDGMENTS The authors are grateful for the technical assistance provided by T. S. Tuladhar and J. Kuikel of the Public Health Section of the Central Health Laboratory, Kathmandu, Nepal, and to Betty Jane Ogg for both technical and secretarial assistance.
LiTERATURE CITED 1. Adams, M. H. 1959. Bacteriophage, p. 450-451. Interscience Publishers, New York. 2. Finkelstein, R. A. 1975. Immunology of cholera. Curr. Top. Microbiol. Immunol. 67:137-195. 3. Finkelstein, R. A., and S. Mukerjee. 1963. Hemagglutination: a rapid method for differentiating Vibrio cholerae and El Tor vibrios. Proc. Soc. Exp. Biol. Med. 112:355-359. 4. Gallut, J., and P. Nicolle. 1963. Lysogenie et lysotypie de V. cholerae et V. El Tor d'origines geographiques diverses. Bull. W.H.O. 28:389-393. 5. Gan, K. M., and S. K. Tjia. 1963. A new method for the differentiation of Vibrio comma and Vibrio eltor. Am. J. Hyg. 77:184-186. 6. Gangarosa, E. J., A. Sanati, H. Saghari, and J. E. Feeley. 1967. Multiple serotypes of Vibrio cholerae from a case of cholera. Evidence suggesting in vivo mutation. Lancet 1:646-648. 7. Newman, F. S., and A. Eisenstark. 1964. Phage host relationships in Vibrio cholerae. J. Infect. Dis. 114:217-225. 8. Ogg, J. E., and W. Chang. 1972. Phage conversion of serotypes in Vibrio fetus. Am. J. Vet. Res. 33:1023-1029. 9. Parker, C., S. N. Richardson, and W. R. Romig. 1970. Production of bacteriophage-associated materials by Vibrio cholerae: possible correlation with pathogenicity. Infect. Immun. 1:417-420. 10. Pollitzer, R. 1959. Cholera. W.H.O. Monogr. Ser. 43:836-841. 11. Rizvi, S. S. IL, and K. A. Monsur. 1965. Observations on the interrelationship of choleraphages and El Tor vibrios, p. 19-24. In Proceedings of the Cholera Research Symposium, 24-29 January 1965. U. S. Government Printing Office, Washington, D.C. 12. Sack, R. B., and C. E. Miller. 1969. Progressive changes in vibrio serotypes in germ-free mice infected with Vibrio cholerae. J. Bacteriol. 99:688-695. 13. Sakazaki, R., and K. Tamura. 1971. Somatic antigen variation in Vibrio cholerae. Jpn. J. Med. Sci. Biol. 24:93-100. 14. Shewan, J. M., and M. Veron. 1974. Genus I. Vibrio Pacini 1854, 168, p. 341-345. In R. E. Buchanan and N. E. Gibbons (ed.), Bergey's manual of determinative bacteriology, 8th ed. The Williams & Wilkins Co., Baltimore. 15. Shrivastava, D. L. 1965. Immunochemistry of vibrio antigens, p. 244-247. In Proceedings of the Cholera
238 16. 17.
18. 19.
OGG, SHRESTHA, AND POUDAYL
Research Symposium, 24-29 January 1965. U. S. Government Printing Office, Washington, D.C. Shrivastava, D. L, and White, P. B. 1947. Note on the relationship of the so-called Ogawa and Inaba types of V. chokerae. Indian J. Med. Res. 35:117-129. Smith, H. L., and K. Goodner. 1965. On the classification of vibrios, p. 4-8. In Proceedings of the Cholera Research Symposium, 24-29 January 1965. U. S. Government Printing Office, Washington, D.C. Takeya, K., and S. Shimodori. 1963. "Prophage-typing" of El Tor vibrios. J. Bacteriol. 85:957-958. Takeya, K., Y. Zinnaka, S. Shimodori, Y. Nakayama,
INFECT. IMMUN. K. Amako, and K. Iida. 1965. Lysogeny in El Tor vibrios, p. 24-29. In Proceedings of the Cholera Research Symposium, 24-29 January 1965. U. S. Government Printing Office, Washington, D.C. 20. Uetake, H., S. E. Luria, and J. W. Burrows. 1958. Conversion of somatic antigens in Salmonella by phage infection leading to lysis or lysogeny. Virology 5:68-91. 21. Vieu, J. F., P. Nicolle, and J. Gallut. 1965. Electron microscopy of some cholera phages, p. 34-36. In Proceedings of the Cholera Research Symposium, 24-29 January 1965. U. S. Government Printing Office, Washington, D.C.
Vol. 19, No. 1
INFECTION AND IMMUNITY, Jan. 1978, p. 231-238
Printed in U.S.A.
Copyright i) 1978 American Society for Microbiology
Phage-Induced Changes in Vibrio cholerae:. Serotype and Biotype Conversions JAMES E. OGG,l* MADAN B. SHRESTHA,2 AND LAXMAN POUDAYL2 Department of Microbiology, Colorado State University, Fort Collins, Colorado 80523,1 and the Central Health Laboratory, Kathmandu, Nepal' Received for publication 25 May 1977
Phage infection of Vibrio cholerae resulted in antigenic changes. A strain of biotype cholerae serotype Ogawa was converted into serotype Hikojima and gained the ability to synthesize antigenic factor C. Some phage-converted strains remained stable after subculture and were immune to superinfection with the same phage. The stable converted strains were lysogenic and released phage having a host range similar to the phage of the donor strain. Reinfection of unstable converted strains which had "lost" antigen C yielded types able to again synthesize this antigen. The polymyxin resistance character was expressed in the biotype cholerae strain after infection with some phage preparations. These polymyxin-resistant strains possessed three main characteristics of El Tor vibrios. The phage-induced changes described provide V. cholerae with the potential for innumerable genetic combinations.
Lysogeny in Vibrio cholerae biotype eltor was first reported by Newman (Bacteriol. Proc., p. 77, 1960) and later by others (4, 18). The general extent of lysogeny in biotype eltor and the morphological and biological properties of El Tor phages have been investigated (7, 19, 21). A correlation exists between pathogenicity and lysogeny in both the El Tor and classical strains of V. cholerae (9, 18). Newman and Eisenstark (7) suggested that phage conversion may account for the hemolytic properties of biotype eltor; they also found that V. cholerae biotype cholerae may be lysogenized with temperate phage from El Tor strains. In this paper we present evidence of bacteriophage conversion affecting the serotype of V. cholerae biotype cholerae. What appeared to be transduction of the polymyxin resistance character by phage from a biotype eltor to a biotype cholerae also occurred with some phage preparations. Stable lysogenization of the host strain was demonstrated by (i) i,munity of the converted strain to superinfection with the same phage after five single colony subcultures in broth and (ii) production of phage by the converted strains after the fifth subculture, which had a host range similar to the phage from the donor strain. MATERIALS AND METHODS Bacterial strains. The V. cholerae strains were isolated at the Central Health Laboratory, Kathmandu, Nepal, from stool specimens collected from clinical cases of Asiatic Cholera. Stock cultures were
maintained on nutrient agar slants held at room temperature. The main characteristics of the strains are listed in Table 1. The biotype determination was made on the basis of sensitivity to 50 IU of polymyxin B (Difco) using the disk test (5), the Voges-Proskauer reaction performed on cultures grown in MR/VP medium (Oxoid) at 22 and 37°C, and the chicken erythrocyte agglutination slide test (3). Strains designated biotype eltor were not sensitive to polymyxin B, were Voges-Proskauer positive at both 22 and 37°C, and agglutinated chicken erythrocytes. Strain 029, classified as biotype cholerae, was sensitive to polymyxin B, gave a weakly positive Voges-Proskauer reaction at 37°C and a negative Voges-Proskauer test at 22°C, and did not agglutinate chicken erythrocytes. V. cholerae polyvalent Ogawa and Inaba agglutinating serums supplied by Wellcome Reagents Limited, Wellcome Research Laboratories, Beckenham, England, were used routinely in slide agglutination tests to determine the serotype of strains and the phageconverted isolants. Bacto-V. cholerae Inaba, Ogawa, and Hikojima antiserums (Difco) served as a second source of typing sera in both slide and tube agglutination tests to double check the serotypes of the strains listed in Table 1, as well as the serotypes of some of the phage-converted isolants. Cells of Hikojima serotype sometimes exhibited a slight delay in agglutination in polyvalent and specific absorbed Inaba and Ogawa antisera. Media. Nutrient broth (Oxoid) and nutrient agar (Oxoid) at pH 7.4 were used for the cultivation of V. cholerae and for bacteriophage propagation. The medium used for carbohydrate fermentation tests was Andrade peptone water (Oxoid) containing 0.5% carbohydrate. Isolation and propagation of bacteriophage. The two biotype eltor serotype Hikojima strains 1633
231
232
OGG, SHRESTHA, AND POUDAYL
INFECT. IMMUN.
TABLE 1. Characteristics of V. cholerae strains Fermentation (without gas) after 48 h of: Strain
Biotype
Serotype
0 Factors Glucose
Mannose
Sucrose
Lactose
029 cholerae + + + + (72 h) Ogawa AB 1621 eltor + Ogawa + + AB + (96 h) 1633a eltor Hikojima + ABC + + + 2001a eltor Hikojima ABC + + + + a Strains designated as serotype Hikojima, as described in Bergey's Manual (14), agglutinate in both Ogawaand Inaba-absorbed antisera.
and 2001 were selected as phage donors from 24 strains screened for evidence of lysogeny using the cross-lysis technique. Phage was released from all but 4 of the 24 strains, including strains 029 and 1621, using the following methods. Supernatants were collected from broth cultures which were variously (i) 24 h old; (ii) irradiated for 4 min with a Sylvania (G15T8) germicidal lamp at a distance of 45 cm, diluted with 2 volumes of sterile broth, and incubated for 7 h at 370C, and (iii) held at 560C for 1 h, diluted with 2 volumes of broth, and incubated for 7 h at 37°C (11). The supernatants were screened by spotting a lawn of bacteria prepared with cells suspended in 0.6% nutrient agar containing 2 x 102 M CaCl2. Strain 1621 was an indicator for phage from 12 of the 20 lysogenic strains. The phage from lysogen 1621 did not lyse cells of the other three strains listed in Table 1. Single turbid plaques formed by phage from either strain 1633 or 2001 on lawns of strain 1621 were picked, and each was placed in 1 ml of nutrient broth. Ten milliliters of a 7- to 8-h broth culture of strain 1621 was inoculated with 0.1 ml of a single plaque phage suspension and incubated overnight at 370C. The cells were sedimented in a centrifuge, and the supernatant was collected. In one experiment, strain 029 was substituted for 1621 for the isolation and propagation of phage from strains 1633 and 2001. The agar overlay method (1) was used to determine phage titer and to obtain phage preparations of high titer. The phage preparations were filtered through a membrane filter (0.45-,m pore diameter). The phage lysates were tested before use to confirm cell-free preparations by plating on nutrient agar and incubating the plates for 48 h. Procedure for phage conversion experiments. Two methods were used to infect recipient bacteria with phage from donor strains 1633 and 2001. The broth method was used in most experiments and consisted of mixing 0.2 ml of cell-free phage suspension (titer, about 2 x 109 to 3 x 109 plaque-forming units per ml) with 0.8 ml of bacterial cells from an overnight culture incubated at 370C (viable cell count, about 1 x 109 to 4 x 109 cells per ml) in a sterile vial containing 1 ml of nutrient broth. Sterile CaCl2 solution was added to give a final concentration of 2 x 10-2 M. Two controls were prepared: one with recipient bacteria in nutrient broth containing CaCl2 to check for possible spontaneous changes in serotype and one containing phage in nutrient broth as a double check on the phage preparation being cell free. The cells in broth control were treated in the same manner as the phage-treated culture. The phage in nutrient broth
control was incubated at 370C for 24 h, and 0.1 ml was spotted onto a nutrient agar plate, which was then incubated for 48 h. The phage-treated cells were incubated at room temperature for 3 and then for 4 h at 370C. A loopful of phage-treated culture was streaked onto an agar plate. After overnight incubation at 370C, isolated colonies were picked at random, and each was placed in 10 ml of broth. After incubation, a loopful of each culture was spread over the surface of an agar plate. A polymyxin B disk was placed in the center of the inoculum, and the plate was incubated overnight. The agar cultures were read for sensitivity or resistance to polymyxin B, and the serotype(s) was determined. A small amount of growth from agar plate cultures showing either a change in serotype or a change in the polymyxin reaction was inoculated into 10 ml of broth and incubated overnight. A nutrient agar plate was streaked with a loopful of broth culture and incubated. Isolated colonies were streaked onto nutrient agar slants to establish stock cultures from single colony isolants. These stock cultures were coded with a number followed by "TP" to indicate treated with phage. A second method (the plate method) was the use of a soft agar overlay containing 0.2 ml of phage and 0.3 ml of an overnight broth culture mixed in 3 ml of agar containing 2 x 10-2 M CaCl2. After incubation of the plates for 2 h at room temperature and then for 24 h at 370C, isolated colonies from areas of confluent lysis were picked, and each was placed in 10 ml of broth. After incubation, this broth culture was tested for its polymyxin reaction and serotype(s) using the same procedure as in the broth method. Stock cultures of the phage-treated strains were established. Procedure for testing stability of phage-converted characteristics. A loopful of growth was transferred from a stock culture coded TP into 10 ml of nutrient broth. After overnight incubation at 370C, the culture was streaked onto nutrient agar to obtain isolated colonies. A single colony was picked and used to establish another broth subculture. Each subculture was screened for retention of the phage-induced char-
acteristic(s). Test for lysogenicity and immunity of the phage-converted strains. Lysogenicity of the TP phage-converted strains was determined after the fifth broth subculture by two methods: immunity to infection with homologous phage and release of phage with a host range similar to that of phage donor strains. Immunity to homologous phage was determined by cross-streaking nutrient broth cultures across a band of phage from a donor strain on nutrient agar as well
PHAGE-INDUCED CHANGES IN V. CHOLERAE
VOL. 19, 1978
as by spotting a lawn of a TP strain with a donor phage preparation. The supernatant from 24-h broth cultures and the supernatant collected from UV-induced TP strains were tested for the release of phage by spotting lawns of "indicator" strains of bacteria.
RESULTS Strain 029 (V. cholerae biotype cholerae serotype Ogawa), when treated with phage preparations obtained from strain 1633 (V. cholerae biotype eltor serotype Hikojima), exhibited changes in serotype and/or polymyxin resistance (Table 2). Strain 029 was treated with phage released from a heat-induced culture of 1633, and 11 cultures were established from single colonies. Of these, eight exhibited changes: two gained only the polymyxin resistance characteristic, five gained both polymyxin resistance and the ability to agglutinate in Inaba typing serum, and one remained polymyxin sensitive but gained the ability to agglutinate in Inaba typing serum. (Agglutination in Inaba typing serum TABLE 2. Changes induced in V. cholerae biotype cholerae (strain 029) by phage obtained from strain 1633 (biotype eltor) Culture codea
Poly- Agglutination reaction myxin rec
tio b
Serotype
0 Factors
Ogawa Ogawa Hikojima
AB AB ABC
Ogawa Inaba
Controls 029 1621 1633 Isolants
S R R
+ + +
-
+
(BH) + R AB ITPa Ogawa + + S Hikojima ABC 3TP + R AB 5TP Ogawa + R + Hikojima ABC 6, 7TP Isolants (PH) + + R 26-28TP Hikojima ABC Isolants (BU) + R AB Ogawa 8, 9TP + + S lOTP Hikojima ABC + + R Hikojima ABC 1lTP + AR R Ogawa 12TP + R + Hikojima ABC 15TP R + AB Ogawa 16TP Isolants (PU) + + S 29TP Hikojima ABC + + R 3OTP Hikojima ABC + R AB Ogawa 31TP + + S Hikojima ABC 32TP + + R 33TP Hikojima ABC aTP indicates cultures established from single colonies after phage treatment. The phage was propagated on strain 1621. Only isolants exhibiting characteristics different from strain 029 are listed. B, broth method; P, plate method; H, heat treatment of 1633 to release phage; U, treatment of 1633 to release phage. b S, Sensitivity to polymyxin; R, resistance to polymyxin.
233
indicates that the cells have the ability to synthesize antigenic factor C.) Similar results were obtained using a phage preparation from an UVinduced strain of 1633. Twelve of fourteen isolants had undergone changes: five gained only the polymyxin resistance characteristic, four were polymyxin resistant and exhibited antigenic factor C production, and three remained polymyxin sensitive but agglutinated in Inaba typing serum. In a later independent experiment using the broth method and phage released from strain 1633 after heat treatment, six out of ten cultures established from single colonies gained the ability to agglutinate in Inaba typing serum, but none exhibited polymyxin resistance. The frequency of occurrence of the polymyxin resistance characteristic was not determined in these experiments, but the results indicate that the polymyxin resistance character appeared to be transferred independently of antigenic factor C. Significant was the observation that antigenic factor C occurred in high frequency in phagetreated cells. This level of expression of a serotype characteristic suggests that antigenic factor C production is associated with the infection of 029 cells with phage from 1633. The phage-induced changes of selected isolants were subjected to stability testing, and the polymyxin resistance characteristic (pol+) was found to be unstable in most but not all of the TP cultures (Table 3). Of the four isolants tested which had gained only the polymyxin resistance character, one (31TP) retained the property through five subcultures. One isolant (26TP) which had the dual characteristic (C pol+) retained its polymyxin resistance character but lost antigenic factor C on the fifth subculture. The polymyxin resistance character may be transduced by the phage from strain 1621 into strain 029. If so, in most cases the exogenote apparently is not integrated into the endogenote and the polymyxin resistance character does not persist in the unstable heterogenote. The two TP strains (26TP, 31TP) which retained the polymyxin resistance character through a number of subcultures became useful later for studies on biotype changes. A phage-induced characteristic was considered stable if it survived five subcultures. Of the eight TP isolants screened exhibiting the dual characteristic (C pol+), six retained antigenic factor C, even though seven had lost the polymyxin resistance character. Two of the four isolants which had been converted only to Hikojima serotype maintained the characteristic. The production of antigen C appeared to be correlated with the presence of the 1633 phage genome in the converted strains. The loss of antigenic factor C in some TP isolants could have
234
OGG, SHRESTHA, AND POUDAYL
INFECT. IMMUN.
TABLE 3. Stability ofphage-induced characteristics in strain 029 Culture codea
Controls: 029 1633 1621 Isolants: 3TP 7TP 8TP 9TP lOTP 11TP 15TP 16TP 26TP 27TP 28TP 29TP
Phage-induced characteristicsb
Characteristics- after subculture no.: 1
2
NC NC NC
NC NC NC
NC NC NC
4
5
NC NC NC
NC NC NC
C C pol+ C C C C C pol+ pol+ C C C C C C C pol+ C C C C C C C C C C pol+ pol+ C pol+ C pol+ C pol+ C pol+ C pol+ pol+ C pol+ C C C C C C pol+ C pol+ C C C C C C C C C C C pol+ C C C C C 3OTP 31TP pol+ pol+ pol+ pol+ pol+ pol+ C C 32TP C pol+ C C C C C 33TP aTP indicates a culture established from a single colony after treatment of strain 029 with phage from strain 1633. b C, Gain of antigenic factor C; pol+, resistance to polymyxin B. e-, Characteristics the same as strain 029; NC, no change in original characteristics.
been due to an unstable condition and eventual loss of the phage genome in those cultures. Those cells no longer lysogenic for the 1633 phage genome should become susceptible to reinfection with this particular phage. This was confirmed by experiments where two of the TP isolants tested which had lost antigenic factor C on transfer (15TP and 26TP) regained the ability to agglutinate in Inaba typing serum after treatment with phage from a heat-induced culture of 1633. Following this second phage treatment, two out of five of the 15TP cultures and six out of six of the 26TP cultures established by single colony isolation agglutinated in Inaba typing serum. One of the 15TP isolants converted to serotype Hikojima was also resistant to polymyxin B. Those TP cultures which maintained the antigenic factor C on transfer and carry the 1633 phage genome should be lysogenic as well as immune to infection with homologous phage. Table 4 shows the results from spotting phage preparations onto lawns of bacteria. These immunity and lysogeny tests indicate that those strains which had antigenic factor C after five subcultures in broth were immune to infection with phage from strain 1633 (the original phage donor) and also to phage from 029 (the original recipient strain). The phage released by the stable "converted strains" had the same host
range as the phage from strain 1633. The original recipient strain 029 was lysed by the phage released from the stable antigenic factor C convertants, but not by the phage from strains 26TP and 32TP which had lost the ability to produce antigen C. These two isolants had become nonlysogenic for the 1633 phage, but still retained the prophage carried by the original 029 strain. The phage released by these strains as well as by strains 31TP and 029 (neither of which possessed antigenic factor C) lysed only indicator strain 1621. Therefore, those strains converted to Hikojima serotype apparently still have the original 029 prophage as well as the 1633 prophage. The polymyxin resistance character alone does not confer immunity to phage 1633. Isolants 26TP and 31TP, in which the polymyxin resistance character was maintained, were lysed by phage from strain 1633 and the stable antigen C factor convertants. Phage from strain 2001, like phage from strain 1633, had the ability to convert strain 029 as well as TP isolants 15, 26, and 31 to antigen C production (Table 5). The phage used was released from strain 2001 after UV light treatment and propagated on strain 1621. The broth method was used in treating the recipient cells with phage. Phage 2001 was very effective in converting all the test strains to Hikojima sero-
PHAGE-INDUCED CHANGES IN V. CHOLERAE
VOL. 19, 1978
235
TABLE 4. Immunity and lysogeny in various V. cholerae strainsa
source)n source) 029 1633 7TP
chokrae strain used for lawn
V.
Phage
27TP
29TP
(pol+)
(C)
28TP
(C)
3OTP
(C)
(pol+)
32TP
1621b
-
-
-
-
-
-
-
-
-
-
-
+
-
-
-
-
+
+
+ +
+
+
+
029b 029b
1633b 1633b
7TP
-
-
+ +
-
(C)
1OTP
26TP
(C)
(C)
-+
31TP
+ + + + + lOTP + _ _ 26TP + + + + + 27TP + + + + + 28TP + + + + + 29TP + + + _ + + 3OTP + 31TP t 32TP a TP indicates a culture established from a single colony after treatment of 029 with phage from strain 1633. Symbol in parentheses indicates the phage-induced character retained after five subcultures: C, antigenic factor C; pol+, resistance to polymyxin B. +, Lysis; -, no lysis. b Control. -
-
-
-
-
-
-
-
-
-
-
-
-
-
-
TABLE 5. Changes induced in V. cholerae biotype cholerae strains 029 and 15TP and in polymyxinresistant derivatives 26TP and 31TP by phage from strain 2001. Culture code
Poly- Agglutination myxin reactiona
Serotype
0 Factors
Ogawa Ogawa
AB AB AB AB
Ogawa Inaba
Controls' 029 26TP 31TP 15TP Isolants from 26TP
S R R S
+ + + +
-
7OTP 71-74TP Isolants from 31TP 75-77TP Isolants from
R R
+ +
+
Ogawa Hikojima
AB ABC
R
+
+
Hikojima
ABC
S S S
+ +
+ -
Hikojima Ogawa
ABC
+
+
Hikojima
ABC
Ogawa
Ogawa
029
8OTP 81TP
82-88TP
-
AB
Isolants from 15TP + + Hikojima ABC 120-124TP S a S, Sensitivity to polymyxin B; R, resistant to polymyxin B. bStrains 15TP, 26TP, and 31TP were isolated originally from strain 029 subsequent to infection with phage from strain 1633.
type. Again, the frequency of the occurrence is typical of a phage conversion system. No polymyxin-resistant isolants were recovered from the phage-treated culture of 029 or strain 15TP, but relatively few cells were screened. Three of the
four phage-converted strains of 26TP, five of the six tested phage-converted strains of 029 and all five of those from 15TP remained stable after five transfers. Only one of the phage-converted strains of 31TP was tested, and it lost antigen C on the third broth subculture. Phages from 1633 and 2001 propagated on strain 029 had the ability to convert strains 029, 15TP, 26TP, and 31TP to Hikojima serotype. Thus, the strain on which the phage was propagated did not apparently affect the expression of the 1633 or 2001 phage genome in a recipient cell. Also, strain 029, subsequent to infection with phage in a supernatant collected from a UV-induced culture of strain 1633, yielded isolants which agglutinated in both Inaba and Ogawa typing serums. These results indicate the 1633 phage genome caused induction of antigenic factor C production by cells of strain 029. As described previously, phage from strain 1633 apparently transduced the polymyxin resistance character from a biotype eltor into biotype cholerae strain 029. The question arose whether other biotype eltor characteristics were associated with a strain being resistant to polymyxin B. The stable polymyxin-resistant "transductants" of strain 029 were tested for ability to agglutinate chicken erythrocytes and ability to produce acetylmethylcarbinol at 22°C (Table 6). The polymyxin-resistant strains (26TP, 31TP) agglutinated chicken erythrocytes and were Voges-Proskauer positive at 220C. These transductants then possessed three main characteristics that distinguish the biotype eltor from biotype cholerae (14). Some of the phageconverted strains which possessed the polymyxin resistance character on original isolation but lost this character on transfer (strains 7TP,
236
OGG, SHRESTHA, AND POUDAYL
TABLE 6. Voges-Proskauer and chicken erythrocyte agglutination reactions by polymyxin-resistant transductants and other strains of V. cholerae Strain
Controls 029
Chicken
Voges-Proskauer reac-
erythro-
tiona at: 22°C 37°C
cyte agglutination
-
Slightly +
Polymyxin reaction
-
S
+ + +
R R R
in 18 h 1633 2001 1621 Isolants 26TPC
+ + +
+
+
31TPC
+
+
+
R R
7TP
-
-
S
+ + +
+
Slightly + in 2 h
-
Slightly + in
-
S
15TP
-
1h Slightly + in 2h
-
S
27TP 28TP 29TP
-
+
-
S
+
-
Slightly + in
-
S S
30TP
-
Slightly + in
-
S
1OTP
2h
2h a +, Positive for acetylmethylcarbinol;-, negative 24 h after adding the reagents. b S, Sensitivity to polymyxin B; R, resistant to polymyxin B. c
Stable polymyxin-resistant transductant of strain
029.
15TP, 27TP, 28TP, and 3OTP) did not exhibit the distinguishing characteristics of biotype eltor. The stable antigen C convertants (10 and 29TP) also did not demonstrate any biotype eltor reactions.
DISCUSSION Phage conversion leading to changes in the serotype of V. cholerae has been demonstrated. Cells of the converted strains (derived from a serotype Ogawa treated with phage from strains of serotype Hikojima) agglutinated in Inaba as well as Ogawa typing sera, indicating a gain in ability to synthesize antigenic factor C, a typespecific factor of Inaba serotype (14). This system of phage conversion influencing the production of an antigenic factor is similar to that reported in Salmonella (20) and in Campylobacter fetus (formerly designated V. fetus) by Ogg and Chang (8). However, in C. fetus and in some of the phage-converted Salmonella, one of the cell's original antigenic factors was lost or repressed on subculture. The stable converted V. cholerae strains maintained their original type-specific antigens as well as the "new" antigenic factor C after five subcultures in broth.
INFECT. IMMUN.
The data indicate that the loss of antigenic factor C in the unstable phage-converted isolants was associated with a change from the lysogenic to the nonlysogenic state in the cells for the original converting phage (Table 4, strains 26TP, 32TP). The Ogawa-to-Hikojima serotype change was induced in biotype cholerae by phage released from strains 1633 and 2001. Strains 1633 and 2001, even though of the same serotype, were isolated from clinical cases of cholera at different times. The temperate phage released by these two strains may be the same or different, but they have the same effect on the host cell: the induction of antigen C production. This system of phage conversion may help explain some reports of serotype conversion in V. cholerae occurring under various conditions. Sakazaki and Tamura (13) suggested that serotype Inaba originated from Ogawa by loss of somatic antigenic factor B; they also recovered Hikojima variants from Ogawa cultures grown in homologous specific factor serum. Conversions from Ogawa to Inaba and from Inaba to Ogawa were observed by Sack and Miller (12) in mono-contaminated gnotobiotic mice. They suggested that such serotype changes were due to the selective effect of antibody within the lumen of the intestine. Earlier, Shrivastava and White (16) isolated Inaba and rough strains after cultivation of serotype Ogawa in monospecific antiserum, whereas predominantly Inaba-like serotypes (which possessed an imperfect Ogawa factor) yielded Ogawa types in Inaba antiserum. Serotype conversions in cholera patients were reported by Gangarosa et al. (6). Loss or gain of antigenic factors in V. cholerae may be due to the expression of a phage genome(s), the loss of a prophage and the selection of nonlysogenic cells as could occur under various culture conditions and in the presence of type-specific antibody, or the infection of a cell by phage released by a strain having the same or a different serotype. As shown by the results of our experiments, cells which become nonlysogenic for a particular temperate phage may regain competency to synthesize a specific antigenic factor (Table 5, strains 15TP, 26TP). When exposed to type-specific antibody, one may expect that the same type of change from lysogenic to nonlysogenic would occur in some cells of V. cholerae as obtained in Salmonella (20) and C. fetus
(8).
More than one type of prophage may become part of the genetic complement of V. cholerae (Table 4). Strain 029, which is lysogenic for a phage which lyses 1621, also became lysogenic for another type of phage when infected with phage from strain 1633. How many prophages
VOL. 19, 1978
PHAGE-INDUCED CHANGES IN V. CHOLERAE
may become part of the genetic complement of this organism is unknown. But the possible interactions that may occur between phage genes and between prophage and host chromosome genes confer onto V. cholerae the potential for numerous genetic and phenotypic variations. Phage conversion could provide additional information about phage genome effects on the expression of host chromosomal genes in this important bacterial pathogen. The polymyxin resistance character also was exhibited in V. cholerae biotype cholerae subsequent to infection by temperate phage from V. cholerae biotype eltor. Associated with the polymyxin resistance character were two other characteristics typical of El Tor biotype: the ability to agglutinate chicken erythrocytes and production of acetylmethylcarbinol at 220C. Although the polymyxin resistance proved to be highly unstable in the transductants, two did retain the characteristic. Until the last cholera pandemic, El Tor vibrios were usually associated with localized outbreaks of cholera-like disease and were not considered epidemic strains like classical cholera vibrios (2). The pandemic strain of V. cholerae biotype eltor might have originated from a classical cholera vibrio that was transduced to polymyxin resistance with phage from a biotype eltor or it could have resulted from a mutation to polymyxin resistance in a classical vibrio. The predominant serotype of V. cholerae may change from one epidemic to another (10). Shrivastava (15) noted changes in dominant serotypes in Calcutta (Ogawa in 1964 versus Inaba in 1962 and 1963) and postulated that such changes could be due to conversion of the antigenic patterns of vibrio strains in vivo or to possible conversions under in vitro environmental conditions. Local conversions in the strains of vibrio rather than importation from outside could account for strains of different characteristics. In Nepal, fluctuations in the dominant serotype also have been occurring from one period to another (unpublished observations of the authors). Phage conversion as demonstrated in our work could be one source of such antigenic changes as occur in nature, since it is a highly effective system for introducing genetic information into a cell of V. cholerae. It may account for the origin of highly virulent strains which cause explosive and severe epidemics threatening the health of over half of the world's population. Smith and Goodner stated in 1965 (17), "... .we are forced to the supposition that a vibrio may well contain latent information toward several syntheses as regard antigens and enzymes, but we have no understanding of the force which
237
may activate these bits of stored information. One must also consider that in nature the vibrios probably have access to a great amount of information from other sources." We feel that the data presented in our paper identify one such source of genetic information available to the vibrios: the influence of temperate phage genomes on the synthesis of type-specific antigens and their ability to serve as a passive carrier of host genes. It also is well known that the introduction of virus genetic material into a cell can activate the expression of latent genetic information of the host chromosome. ACKNOWLEDGMENTS The authors are grateful for the technical assistance provided by T. S. Tuladhar and J. Kuikel of the Public Health Section of the Central Health Laboratory, Kathmandu, Nepal, and to Betty Jane Ogg for both technical and secretarial assistance.
LiTERATURE CITED 1. Adams, M. H. 1959. Bacteriophage, p. 450-451. Interscience Publishers, New York. 2. Finkelstein, R. A. 1975. Immunology of cholera. Curr. Top. Microbiol. Immunol. 67:137-195. 3. Finkelstein, R. A., and S. Mukerjee. 1963. Hemagglutination: a rapid method for differentiating Vibrio cholerae and El Tor vibrios. Proc. Soc. Exp. Biol. Med. 112:355-359. 4. Gallut, J., and P. Nicolle. 1963. Lysogenie et lysotypie de V. cholerae et V. El Tor d'origines geographiques diverses. Bull. W.H.O. 28:389-393. 5. Gan, K. M., and S. K. Tjia. 1963. A new method for the differentiation of Vibrio comma and Vibrio eltor. Am. J. Hyg. 77:184-186. 6. Gangarosa, E. J., A. Sanati, H. Saghari, and J. E. Feeley. 1967. Multiple serotypes of Vibrio cholerae from a case of cholera. Evidence suggesting in vivo mutation. Lancet 1:646-648. 7. Newman, F. S., and A. Eisenstark. 1964. Phage host relationships in Vibrio cholerae. J. Infect. Dis. 114:217-225. 8. Ogg, J. E., and W. Chang. 1972. Phage conversion of serotypes in Vibrio fetus. Am. J. Vet. Res. 33:1023-1029. 9. Parker, C., S. N. Richardson, and W. R. Romig. 1970. Production of bacteriophage-associated materials by Vibrio cholerae: possible correlation with pathogenicity. Infect. Immun. 1:417-420. 10. Pollitzer, R. 1959. Cholera. W.H.O. Monogr. Ser. 43:836-841. 11. Rizvi, S. S. IL, and K. A. Monsur. 1965. Observations on the interrelationship of choleraphages and El Tor vibrios, p. 19-24. In Proceedings of the Cholera Research Symposium, 24-29 January 1965. U. S. Government Printing Office, Washington, D.C. 12. Sack, R. B., and C. E. Miller. 1969. Progressive changes in vibrio serotypes in germ-free mice infected with Vibrio cholerae. J. Bacteriol. 99:688-695. 13. Sakazaki, R., and K. Tamura. 1971. Somatic antigen variation in Vibrio cholerae. Jpn. J. Med. Sci. Biol. 24:93-100. 14. Shewan, J. M., and M. Veron. 1974. Genus I. Vibrio Pacini 1854, 168, p. 341-345. In R. E. Buchanan and N. E. Gibbons (ed.), Bergey's manual of determinative bacteriology, 8th ed. The Williams & Wilkins Co., Baltimore. 15. Shrivastava, D. L. 1965. Immunochemistry of vibrio antigens, p. 244-247. In Proceedings of the Cholera
238 16. 17.
18. 19.
OGG, SHRESTHA, AND POUDAYL
Research Symposium, 24-29 January 1965. U. S. Government Printing Office, Washington, D.C. Shrivastava, D. L, and White, P. B. 1947. Note on the relationship of the so-called Ogawa and Inaba types of V. chokerae. Indian J. Med. Res. 35:117-129. Smith, H. L., and K. Goodner. 1965. On the classification of vibrios, p. 4-8. In Proceedings of the Cholera Research Symposium, 24-29 January 1965. U. S. Government Printing Office, Washington, D.C. Takeya, K., and S. Shimodori. 1963. "Prophage-typing" of El Tor vibrios. J. Bacteriol. 85:957-958. Takeya, K., Y. Zinnaka, S. Shimodori, Y. Nakayama,
INFECT. IMMUN. K. Amako, and K. Iida. 1965. Lysogeny in El Tor vibrios, p. 24-29. In Proceedings of the Cholera Research Symposium, 24-29 January 1965. U. S. Government Printing Office, Washington, D.C. 20. Uetake, H., S. E. Luria, and J. W. Burrows. 1958. Conversion of somatic antigens in Salmonella by phage infection leading to lysis or lysogeny. Virology 5:68-91. 21. Vieu, J. F., P. Nicolle, and J. Gallut. 1965. Electron microscopy of some cholera phages, p. 34-36. In Proceedings of the Cholera Research Symposium, 24-29 January 1965. U. S. Government Printing Office, Washington, D.C.
