THE JOURNAL OF INFECTIOUS DISEASES • VOL. 133, NO.3· © 1976 by the University of Chicago. All rights reserved.
MARCH 1976
MAJOR ARTICLES Klebsiella L-Forms: Effect of Growth as L-Form on Virulence of Reverted Klebsiella pneumoniae L. B. Guze, H. J. Harwick, * and G. M. Kalmanson
From the Research and Medical Services, Veterans Administration, Wadsworth Hospital Center, Los Angeles, California; the UCLA School of Medicine, Los Angeles; and Harbor General Hospital, Torrance, California
Among possible mechanisms by which L-forms may be implicated in the pathogenesis of disease is in vivo reversion to the parental classic bacterial form. For this mechanism to be workable, the reverted bacteria would have to retain their original virulence. The present study was directed to this point. A mouse-virulent strain of Klebsiella pneumoniae was serially passaged in vitro as a penicillin-induced L-form. At various intervals it was reverted to the bacterial form, and virulence was reevaluated. With an increasing number of passages as L-forms, the revertants gradually lost
their virulence and eventually became completely avirulent. Studies were then done to elucidate the factor(s) responsible for this loss of virulence and the physiologic and biochemical characteristics that were different in revertant and parental forms. Other studies demonstrated that the avirulent revertant immunized mice against infection with the virulent parent. Materials and Methods
Bacteria. K. pneumoniae laboratory strain no. 163, type 1 (ATCC 8045, American Type Culture Collection, Rockville, Md.) , originally obtained from human blood culture, was used in these studies. Strain no. 276, serotype 2, obtained from a blood culture, was used for heterologous challenge in immunization experiments. Production, propagation, and reversion 0/ Lforms. Bacteria were inoculated into brain-heart infusion broth (BHI; Difco, Detroit, Mich.), incubated overnight, and plated on brain-heart infusion agar (BHIA; Difco) containing 2 % bovine serum albumin (BSA, fraction V; Armour, Chicago, III.) with 1. 8 % NaCl added as osmotic
Received for publication January 6, 1975, and in revised form July 31, 1975. This investigation was supported by project no. 3498-01 from the Veterans Administration. We thank Mr. J. Kaszlowski, Mr. K. Ishida, Mr. R. Wisell, Mr. N. Liu, and Ms. J. Hardin for their excellent technical assistance, and Dr. M. Fox (Department of Mathematics, UCLA) for his statistical help. Please address requests for reprints to Dr. Lucien B. Guze, Veterans Administration, Wadsworth Hospital Center, Wilshire and Sawtelle Boulevards, Los Angeles, California 90073. * Deceased.
245
Downloaded from http://jid.oxfordjournals.org/ at University of Birmingham on August 31, 2015
A strain of Klebsiella pneumoniae virulent for mice was serially passaged in vitro as a penicillin-induced L-form. Periodically, the L-form was reverted to the bac~ terial form by removal of penicillin, and the revertant was tested for virulence. Virulence was gradually reduced, and after 109 passages the revertant was totally avirulent. Virulence was not restored by 40 passages of the revertant through mouse peritoneum. The revertant grew less vigorously and did not infect mice when inoculated subcutaneously; in contrast, the parental form was found in all tissues examined. After phagocytosis in vitro or in vivo, the revertant survived but did not increase in numbers, whereas the parental form did increase in numbers. The revertant had a much smaller capsule than did the parental form. Immunization with the live revertant resulted in type-specific protection against infection with the parental form.
246
Growth rate was measured both in BHI broth and in Hank's solution (0.15% glucose, 0.02% BSA, and 2.5% normal mouse serum) with an initial concentration of about 105 organisms/ml. The bacterial population was quantitated by the pourplate procedure at intervals up to 24 hr. The size of bacteria was estimated in two ways. The first method was direct measurement of an India ink preparation with a calibrated eyepiece and a X 100 objective. The second method was the preparation of cell envelopes from a known number of bacteria according to the method of Williams and Taylor-Robinson [2]. Sensitivity to antibiotic disks was tested according to standard methods. Mueller-Hinton broth and agar were used. Chemotactic activity was studied in modified Boyden chambers [3, 4]. Leukocytes were obtained from mouse blood by sedimentation with dextran. After incubation for 3 hr, the chemotactic index was calculated by division of the number of leukocytes on the bottom of the filter by the number on the top. Thus, the larger the index, the greater the level of chemotaxis. Invasive properties. Invasiveness was determined by sc inoculation of the 18-hr revertant and parent klebsiellae in suitable dilutions. At 24 hr mice were sacrificed, and the bacterial populations in kidney, liver, spleen, and blood of each animal were quantitated by plate count. A swab culture of the peritoneum was made on BHIA to determine whether any bacteria were present. Phagocytosis in mouse peritoneum. Eighteenhour cultures of the bacteria were washed twice and resuspended in Hanks' solution to a concentration of 109 organisms/ml. Twelve mice were given an injection of 108 bacteria. At 0, 1, 2, and 4 hr, three mice were anesthetized with ether, the peritoneum was exposed, and Hanks' solution (2 ml) was injected, mixed, and withdrawn. A 0.1ml portion was diluted in 2 ml of the same diluent and homogenized, and quantitative bacterial assays were done. The remaining fluid was spun at 210 g for 5 min, and Wright-stained smears were made from the sediment. In one experiment the method was modified slightly to permit estimation of the viability of ingested bacteria. The peritoneal fluid was divided into two portions: one portion was used to make the smear, and the other was gently centrifuged to sediment the
Downloaded from http://jid.oxfordjournals.org/ at University of Birmingham on August 31, 2015
stabilizer (BHIA-BSA-NaCI) and (in some cases) 3,125-6,250 units of penicillin G/ml. After incubation for two to three days at 37 C, a segment of agar containing two or three typical L-form colonies was cut out and homogenized with 2 ml of BHI-BSA-NaCI broth. The homogenate was transferred to BHI-BSA-NaCI-penicillin plates and labeled T1 (transfer no. 1). Serial transfers were done similarly twice a week. The yield of L-forms increased, and it was found that the amount of penicillin had to be increased gradually to a level of 30,000 units/ml to prevent spontaneous reversion to the parental bacterial form. Planned reversion to the parental bacterial form was accomplished by inoculation of the organism into BHI-BSA-NaCI broth without peni.. cillin. Reversion occurred after one or two transfers without penicillin. Virulence w,as lost completely after 109 transfers, and the avirulent revertant was designated RLT 109. The parental bacterial form was transferred twice a week (109 times) on BHIA-BSA-NaCI to determine the effect of prolonged passage on artificial medium. This form remained fully virulent and was designated PT109. RLTI09 and PT109 were used in all comparative studies. Determination of virulence. Randomly bred Swiss mice were obtained from commercial sources. The ip LD50 was determined by the method of Reed and Muench [1] one, two, three, and four days after inoculation of bacteria; 10 mice were given each serial 10-fold dilution. Restoration of virulence. The restoration of virulence of RLT 109 was attempted by passage of the organism through mice. Two mice were inoculated ip with 0.5 ml of a BHI culture incubated overnight. At 4-6 hr one mouse was sacrificed, and the peritoneum was opened, swabbed with a sterile applicator, and streaked onto a BHIA plate. After incubation overnight the plate was stored for five days at 4 C. The next broth culture was prepared from this plate. The other mouse was observed for one week before it was considered to have survived. The whole procedure was repeated serially each week. Serologic, physiologic, and biochemical properties. Serotyping, motility, and a variety of biochemical activities were tested by standard bacteriologic methods.
Guze, Harwick, and Kalmanson
247
Klebsiella L-Form Revertant Virulence
107 ) on the same schedule. The LD50 was determined one week later by challenge with heterologous strain no. 276 (type 2). Results
Effect of L-form growth on virulence of the revertant form. The results of studies of virulence are summarized in figure 1. PT109 retained its virulence, in contrast to the parental form that was not serially transferred on artificial medium. The organism that reverted after one passage as an L-form was similarly virulent. From then on, with the exception of the 20th transfer (T20), there was a gradual reduction in virulence; this reduction became statistically significant at T30T40 and marked by T70 and T90. After 109 transfers the revertant had completely lost its virulence; no deaths resulted from the maximal number of bacteria injected (log 7.32). Virulence of the revertant was not restored by 40 passages through mouse peritoneum. Serologic, physiologic, and biochemical properties. Both the parent and the revertant were type 1. During growth a slight tendency to greater slime formation was noted in the revertant. Neither form was motile. Biochemical studies demonstrated no difference between the two
NO DEATHS
,-----------_TI09
7.00
6.00 ~
5.00
8
:~
~---.
-----------
..J
....~
.T90
T 10
4.00
UJ
-J
~
~
(!)
9
3.00 2.00
1.00
i
»
i
234 DAYS POST INFECTION
Figure 1. Effect of the number of transfers as an L-form on the virulence of reverted Klebsiella pneu.. moniae. To = no transfers, T 5 = five transfers, etc.
Downloaded from http://jid.oxfordjournals.org/ at University of Birmingham on August 31, 2015
phagocytes but not the uningested bacteria. The sedimented leukocytes were resuspended in the original volume and homogenized, and bacteria were quantitated. Surface phagocytosis. The method of Medearis et al. [5] was used, with one exception [6], i.e., beef extract (0.8 ml of a 20% solution) was injected ip to induce a polymorphonuclear response in the peritoneum 18-20 hr later. Tube phagocytosis. To determine the rates of killing of bacteria and of phagocytosis, a tube phagocytosis procedure was used. The method was that described in a previous study of human phagocytosis [7] except that mouse peritoneal leukocytes were obtained as described above. The final concentration of bacteria in Hanks' solution was 5 X 106 /ml, and that of leukocytes was 106 / m!. At zero-time and after incubation for 1, 2, and 3 hr at 37 C, the reaction mixture was divided into three parts. One part was homogenized to disrupt white· cells, and the viable bacteria were quantitatively assayed by plate count. The second part was spun gently (210 g) so that leukocytes were sedimented but bacteria were not freed. The sediment was resuspended to the original volume with distilled water and homogenized, and viable bacteria were assayed by plate count. The third portion was similarly centrifuged, Wright-stained smears were made from the sediment, and the degree of phagocytosis was estimated. Serum bactericidal susceptibility. Bacteria in either the logarithmic or the stationary phase (growth for 18 hr), in a final dilution of 103 organisms/ml, were mixed with serial dilutions of serum. The test was done with mouse serum and (because of known problems involved in the demonstration of serum bactericidal activity with mouse serum) with normal human serum, both with and without guinea pig complement (1: 25) . Surviving bacteria were quantitated by plate count after incubation for 1 hr at 37 C. The diluent used was 0.5% peptone (Difco) with NaCI added to an osmolality of 100 mosmol/kg. Immunization by RLTI09. Mice were immunized by sc inoculation of 1.0-2.8 X 108 live bacteria three times a week for four weeks, and the LD50 was determined one week later with PT 109. In a second experiment, mice were immunized with either live RLT109 (0.9-3.4 X 108 bacteria) or heat-killed RLT109 (3.95 X
248
Figure 2. Growth of parent and revertant Klebsiella pneumoniae in brain-heart infusion broth.
A ntibiotic sensitivity. In general sensitivity and resistance, the parent and the revertant were similar. PT109 and RLT109 were equally sensitive to methacycline, tetracycline, chlortetracycline, oxytetracycline, demethylchlortetracycline, chloramphenicol, ampicillin, penicillin, cephalothin, carbenicillin, polymyxin B, colistin, gentamicin, kanamycin, neomycin, dihydrostreptomycin, streptomycin, rifampin, nitrofurantoin, and novobiocin. Both forms were resistant to erythromycin, lincomycin, methicillin, oxacillin, and vancomycin. PT109 was resistant to bacitracin, while RLTI09 was susceptible. Chemotaxis. PTI09 was distinctly more chemotactic (index, 32) than RLT 109 (index, 18.6). Tissue invasion. When as many as 3.0 X 108 RLT 109 were inoculated sc, no organisms were found in any of the tissues examined. In contrast, 100,000 fewerPT 109 organisms invaded the body, and organisms were found in all tissues examined. Phagocytosis in mouse peritoneum. The fate of bacteria in the mouse peritoneum is shown in figure 3. The data represent an average of four experiments, with three animals for each period (except for results on the fate of phagocytized bacteria, which are derived from one of the experiments). When the whole suspension was considered, counts of the parental form increased progressively, while counts of the revertant form decreased with time. When the fate of phagocytized bacteria was examined alone, both parental and revertant forms increased in number for the first 2 hr. The parental form then continued to increase in number, while the number of viable revertant bacteria declined markedly. Surface phagocytosis. No significant difference between the proportions of leukocytes with ingested bacteria or in numbers of bacteria ingested per leukocyte was noted for revertant and parent bacteria. With both PTI09 and RLTI09, logarithmic growth-phase bacteria were more easily phagocytized than organisms in the stationary phase. Tube phagocytosis. Logarithmic-phase organisms were used in an experiment designed for the study of both phagocytosis and the fate of the ingested organisms. The results (average of three experiments) are shown in figure 4. The parent
Downloaded from http://jid.oxfordjournals.org/ at University of Birmingham on August 31, 2015
forms. The biochemical tests that had positive results were: Kligler iron agar; Voges-Proskauer; nitrite reduction; and fermentation of mannitol, xylose, arabinose, raffinose, sorbitol, sucrose, glucose, lactose, mannose, galactose, maltose, trehalose, salicin, and levulose. Oxidase, indole, and ornithine reactions were negative. Inulin, dulcitol, and glycerol were not fermented. Litmus milk was acidified but not coagulated. Gelatin was not liquefied, H 2 S was not produced, and urease activity was not demonstrated. Growth of the revertant in BHI broth (figure 2) was distinctly slower than that of the parent during the early phase, although both forms reached the same level by 18 hr. Difference in the rate of growth was even more striking when the organisms were tested in Hanks' solution. The parental form grew somewhat more slowly in Hanks' solution than in BHI broth and reached a level of only 107 Iml in Hanks' solution as compared with 109 /ml in BHI broth. However, the revertant did not grow at all in Hanks' solution. Size of bacteria. The average of the transverse diameters of 20 individual PT109 organisms was 4.1 11m, including the capsule; for RLT109 organisms this value was 1.2 11m. Both types of bacterium without capsules had diameters of 1.0 11 m . Weights of cell envelopes were as follows: 2.3 X 1010 parent bacteria yielded 1.95 mg; 2.7 X 1010 revertant bacteria yielded 0.61 mg. This finding indicated that the cell envelope was about 3.5 times larger in the parental form than in the revertant.
Guze, Harwick, and Kalmanson
249
Klebsiella L-Form Revertant Virulence
400
~
ffi
300
c
I-
~
ll..
o
...J
~ 200
~ ::l In
I-
Z
W U
a:::
~ 100
Discussion
0 ......-e5=t=:==------r---.-----r2 3 4 TIME
HOURS
Figure 3. Phagocytosis and survival of parent and revertant Klebsiella pneumoniae in mouse peritoneum. (0--0) = revertant in whole suspension; (0- - -0) = parent in whole suspension; (0--0) = phagocytized parent organisms only; (0- - -0) = phagocytized revertant organisms only; (e--e) = percentage of parent organisms phagocytized; (e- - -e) = percentage of revertant organisms phagocytized.
Prolonged cultivation as a penicillin-induced Lform and periodic reversion to the bacterial form resulted in a step-by-step reduction in virulence of an originally virulent strain of K. pneumoniae in the mouse. After 109 L-form passages, virulence was totally lost and was not restored by 40 passages through mouse peritoneum. This loss
,
I
3000 f
oct
b «
ilID
f
:E: ~
"
~
I
ii:
~ 2500
,
CD
grew vigorously in the diluent without any leuko~ cytes present, while the revertant did not grow. The viable count of parent bacteria in the mixture increased more rapidly than did the count in the control culture, as if some growth-stimulating effect occurred as a result of the presence of leukocytes. On the other hand, the count of viable revertants did not change with time. The parental form obviously grew after phagocytosis, while the revertant did not. At each interval there was more phagocytosis of PTI09 than of RLTI09. Thus the parental form was phagocytized more efficiently than the revertant, but, once engulfed, the parent actually grew within the leukocyte, while the revertant merely survived during the period studied. Bactericidal susceptibility. There was no evidence of killing of either logarithmic- or stationary-phase parent or revertant forms by human or mouse serum after incubation for 1 hr.
~ !AI
f
,f
f
~ 2000
I
...J
~
I
I
f
~ l-
>-
v
... ~ ... e
/ ... / ' " ,.....
~ 1500 :::>
,e/,'/ ,'/ "" /'
~
/'
/,,,
~
500
"""
e ;'
",/' .../.d'
o
I
/'
60
d ::l
~
40 ~
~
/
_---_-D- _------D
100.pf1ftff-;:----
~ ~
...",,,,/ ",.e~ ;f ~
~ 1000
8
::l II::
~I
en
80
_-«/
~
~
2
3
~
20
Z
I-
~
TIME - HOURS
Figure 4. Phagocytosis and survival of parent and revertant Klebsiella pneumoniae in vitro. (x--x) = revertant in control medium (diluent without leukocytes); (x- - -x) = parent in control medium; (0--0) revertant in whole suspension; (0- - -0) parent in whole suspension; (0--0) = phagocytized revertant organisms only; (0- - -0) = phagocytized parent organisms only; (e--e) = percentage of revertant organisms phagocytized; (e- - -e) = percentage of parent organisms phagocytized.
Downloaded from http://jid.oxfordjournals.org/ at University of Birmingham on August 31, 2015
Immunization with RLT109 and challenge with PT109. The LD50 was < 10 bacteria in unimmunized controls. In contrast, the LD50 in immunized mice was 1.08 X 105 bacteria, a value indicating a high degree of protection (figure 5). Immunization with RLT109 and challenge with heterologous strain no. 276. The outcome of inoculation of RLTI09 is shown in figure 6. The LD50 was 170 organisms in unimmunized control mice, 610 organisms in those immunized with heat-killed RLTI09, and 40 organisms in those immunized with live RLT 109. These differences were not statistically significant.
Guze, Harwick, and Kalmanson
250
6.00
au
IMMUNIZED WITH
g
REVERTED.
(/)
-.
LIVE
..J
MARCH 1976
MAJOR ARTICLES Klebsiella L-Forms: Effect of Growth as L-Form on Virulence of Reverted Klebsiella pneumoniae L. B. Guze, H. J. Harwick, * and G. M. Kalmanson
From the Research and Medical Services, Veterans Administration, Wadsworth Hospital Center, Los Angeles, California; the UCLA School of Medicine, Los Angeles; and Harbor General Hospital, Torrance, California
Among possible mechanisms by which L-forms may be implicated in the pathogenesis of disease is in vivo reversion to the parental classic bacterial form. For this mechanism to be workable, the reverted bacteria would have to retain their original virulence. The present study was directed to this point. A mouse-virulent strain of Klebsiella pneumoniae was serially passaged in vitro as a penicillin-induced L-form. At various intervals it was reverted to the bacterial form, and virulence was reevaluated. With an increasing number of passages as L-forms, the revertants gradually lost
their virulence and eventually became completely avirulent. Studies were then done to elucidate the factor(s) responsible for this loss of virulence and the physiologic and biochemical characteristics that were different in revertant and parental forms. Other studies demonstrated that the avirulent revertant immunized mice against infection with the virulent parent. Materials and Methods
Bacteria. K. pneumoniae laboratory strain no. 163, type 1 (ATCC 8045, American Type Culture Collection, Rockville, Md.) , originally obtained from human blood culture, was used in these studies. Strain no. 276, serotype 2, obtained from a blood culture, was used for heterologous challenge in immunization experiments. Production, propagation, and reversion 0/ Lforms. Bacteria were inoculated into brain-heart infusion broth (BHI; Difco, Detroit, Mich.), incubated overnight, and plated on brain-heart infusion agar (BHIA; Difco) containing 2 % bovine serum albumin (BSA, fraction V; Armour, Chicago, III.) with 1. 8 % NaCl added as osmotic
Received for publication January 6, 1975, and in revised form July 31, 1975. This investigation was supported by project no. 3498-01 from the Veterans Administration. We thank Mr. J. Kaszlowski, Mr. K. Ishida, Mr. R. Wisell, Mr. N. Liu, and Ms. J. Hardin for their excellent technical assistance, and Dr. M. Fox (Department of Mathematics, UCLA) for his statistical help. Please address requests for reprints to Dr. Lucien B. Guze, Veterans Administration, Wadsworth Hospital Center, Wilshire and Sawtelle Boulevards, Los Angeles, California 90073. * Deceased.
245
Downloaded from http://jid.oxfordjournals.org/ at University of Birmingham on August 31, 2015
A strain of Klebsiella pneumoniae virulent for mice was serially passaged in vitro as a penicillin-induced L-form. Periodically, the L-form was reverted to the bac~ terial form by removal of penicillin, and the revertant was tested for virulence. Virulence was gradually reduced, and after 109 passages the revertant was totally avirulent. Virulence was not restored by 40 passages of the revertant through mouse peritoneum. The revertant grew less vigorously and did not infect mice when inoculated subcutaneously; in contrast, the parental form was found in all tissues examined. After phagocytosis in vitro or in vivo, the revertant survived but did not increase in numbers, whereas the parental form did increase in numbers. The revertant had a much smaller capsule than did the parental form. Immunization with the live revertant resulted in type-specific protection against infection with the parental form.
246
Growth rate was measured both in BHI broth and in Hank's solution (0.15% glucose, 0.02% BSA, and 2.5% normal mouse serum) with an initial concentration of about 105 organisms/ml. The bacterial population was quantitated by the pourplate procedure at intervals up to 24 hr. The size of bacteria was estimated in two ways. The first method was direct measurement of an India ink preparation with a calibrated eyepiece and a X 100 objective. The second method was the preparation of cell envelopes from a known number of bacteria according to the method of Williams and Taylor-Robinson [2]. Sensitivity to antibiotic disks was tested according to standard methods. Mueller-Hinton broth and agar were used. Chemotactic activity was studied in modified Boyden chambers [3, 4]. Leukocytes were obtained from mouse blood by sedimentation with dextran. After incubation for 3 hr, the chemotactic index was calculated by division of the number of leukocytes on the bottom of the filter by the number on the top. Thus, the larger the index, the greater the level of chemotaxis. Invasive properties. Invasiveness was determined by sc inoculation of the 18-hr revertant and parent klebsiellae in suitable dilutions. At 24 hr mice were sacrificed, and the bacterial populations in kidney, liver, spleen, and blood of each animal were quantitated by plate count. A swab culture of the peritoneum was made on BHIA to determine whether any bacteria were present. Phagocytosis in mouse peritoneum. Eighteenhour cultures of the bacteria were washed twice and resuspended in Hanks' solution to a concentration of 109 organisms/ml. Twelve mice were given an injection of 108 bacteria. At 0, 1, 2, and 4 hr, three mice were anesthetized with ether, the peritoneum was exposed, and Hanks' solution (2 ml) was injected, mixed, and withdrawn. A 0.1ml portion was diluted in 2 ml of the same diluent and homogenized, and quantitative bacterial assays were done. The remaining fluid was spun at 210 g for 5 min, and Wright-stained smears were made from the sediment. In one experiment the method was modified slightly to permit estimation of the viability of ingested bacteria. The peritoneal fluid was divided into two portions: one portion was used to make the smear, and the other was gently centrifuged to sediment the
Downloaded from http://jid.oxfordjournals.org/ at University of Birmingham on August 31, 2015
stabilizer (BHIA-BSA-NaCI) and (in some cases) 3,125-6,250 units of penicillin G/ml. After incubation for two to three days at 37 C, a segment of agar containing two or three typical L-form colonies was cut out and homogenized with 2 ml of BHI-BSA-NaCI broth. The homogenate was transferred to BHI-BSA-NaCI-penicillin plates and labeled T1 (transfer no. 1). Serial transfers were done similarly twice a week. The yield of L-forms increased, and it was found that the amount of penicillin had to be increased gradually to a level of 30,000 units/ml to prevent spontaneous reversion to the parental bacterial form. Planned reversion to the parental bacterial form was accomplished by inoculation of the organism into BHI-BSA-NaCI broth without peni.. cillin. Reversion occurred after one or two transfers without penicillin. Virulence w,as lost completely after 109 transfers, and the avirulent revertant was designated RLT 109. The parental bacterial form was transferred twice a week (109 times) on BHIA-BSA-NaCI to determine the effect of prolonged passage on artificial medium. This form remained fully virulent and was designated PT109. RLTI09 and PT109 were used in all comparative studies. Determination of virulence. Randomly bred Swiss mice were obtained from commercial sources. The ip LD50 was determined by the method of Reed and Muench [1] one, two, three, and four days after inoculation of bacteria; 10 mice were given each serial 10-fold dilution. Restoration of virulence. The restoration of virulence of RLT 109 was attempted by passage of the organism through mice. Two mice were inoculated ip with 0.5 ml of a BHI culture incubated overnight. At 4-6 hr one mouse was sacrificed, and the peritoneum was opened, swabbed with a sterile applicator, and streaked onto a BHIA plate. After incubation overnight the plate was stored for five days at 4 C. The next broth culture was prepared from this plate. The other mouse was observed for one week before it was considered to have survived. The whole procedure was repeated serially each week. Serologic, physiologic, and biochemical properties. Serotyping, motility, and a variety of biochemical activities were tested by standard bacteriologic methods.
Guze, Harwick, and Kalmanson
247
Klebsiella L-Form Revertant Virulence
107 ) on the same schedule. The LD50 was determined one week later by challenge with heterologous strain no. 276 (type 2). Results
Effect of L-form growth on virulence of the revertant form. The results of studies of virulence are summarized in figure 1. PT109 retained its virulence, in contrast to the parental form that was not serially transferred on artificial medium. The organism that reverted after one passage as an L-form was similarly virulent. From then on, with the exception of the 20th transfer (T20), there was a gradual reduction in virulence; this reduction became statistically significant at T30T40 and marked by T70 and T90. After 109 transfers the revertant had completely lost its virulence; no deaths resulted from the maximal number of bacteria injected (log 7.32). Virulence of the revertant was not restored by 40 passages through mouse peritoneum. Serologic, physiologic, and biochemical properties. Both the parent and the revertant were type 1. During growth a slight tendency to greater slime formation was noted in the revertant. Neither form was motile. Biochemical studies demonstrated no difference between the two
NO DEATHS
,-----------_TI09
7.00
6.00 ~
5.00
8
:~
~---.
-----------
..J
....~
.T90
T 10
4.00
UJ
-J
~
~
(!)
9
3.00 2.00
1.00
i
»
i
234 DAYS POST INFECTION
Figure 1. Effect of the number of transfers as an L-form on the virulence of reverted Klebsiella pneu.. moniae. To = no transfers, T 5 = five transfers, etc.
Downloaded from http://jid.oxfordjournals.org/ at University of Birmingham on August 31, 2015
phagocytes but not the uningested bacteria. The sedimented leukocytes were resuspended in the original volume and homogenized, and bacteria were quantitated. Surface phagocytosis. The method of Medearis et al. [5] was used, with one exception [6], i.e., beef extract (0.8 ml of a 20% solution) was injected ip to induce a polymorphonuclear response in the peritoneum 18-20 hr later. Tube phagocytosis. To determine the rates of killing of bacteria and of phagocytosis, a tube phagocytosis procedure was used. The method was that described in a previous study of human phagocytosis [7] except that mouse peritoneal leukocytes were obtained as described above. The final concentration of bacteria in Hanks' solution was 5 X 106 /ml, and that of leukocytes was 106 / m!. At zero-time and after incubation for 1, 2, and 3 hr at 37 C, the reaction mixture was divided into three parts. One part was homogenized to disrupt white· cells, and the viable bacteria were quantitatively assayed by plate count. The second part was spun gently (210 g) so that leukocytes were sedimented but bacteria were not freed. The sediment was resuspended to the original volume with distilled water and homogenized, and viable bacteria were assayed by plate count. The third portion was similarly centrifuged, Wright-stained smears were made from the sediment, and the degree of phagocytosis was estimated. Serum bactericidal susceptibility. Bacteria in either the logarithmic or the stationary phase (growth for 18 hr), in a final dilution of 103 organisms/ml, were mixed with serial dilutions of serum. The test was done with mouse serum and (because of known problems involved in the demonstration of serum bactericidal activity with mouse serum) with normal human serum, both with and without guinea pig complement (1: 25) . Surviving bacteria were quantitated by plate count after incubation for 1 hr at 37 C. The diluent used was 0.5% peptone (Difco) with NaCI added to an osmolality of 100 mosmol/kg. Immunization by RLTI09. Mice were immunized by sc inoculation of 1.0-2.8 X 108 live bacteria three times a week for four weeks, and the LD50 was determined one week later with PT 109. In a second experiment, mice were immunized with either live RLT109 (0.9-3.4 X 108 bacteria) or heat-killed RLT109 (3.95 X
248
Figure 2. Growth of parent and revertant Klebsiella pneumoniae in brain-heart infusion broth.
A ntibiotic sensitivity. In general sensitivity and resistance, the parent and the revertant were similar. PT109 and RLT109 were equally sensitive to methacycline, tetracycline, chlortetracycline, oxytetracycline, demethylchlortetracycline, chloramphenicol, ampicillin, penicillin, cephalothin, carbenicillin, polymyxin B, colistin, gentamicin, kanamycin, neomycin, dihydrostreptomycin, streptomycin, rifampin, nitrofurantoin, and novobiocin. Both forms were resistant to erythromycin, lincomycin, methicillin, oxacillin, and vancomycin. PT109 was resistant to bacitracin, while RLTI09 was susceptible. Chemotaxis. PTI09 was distinctly more chemotactic (index, 32) than RLT 109 (index, 18.6). Tissue invasion. When as many as 3.0 X 108 RLT 109 were inoculated sc, no organisms were found in any of the tissues examined. In contrast, 100,000 fewerPT 109 organisms invaded the body, and organisms were found in all tissues examined. Phagocytosis in mouse peritoneum. The fate of bacteria in the mouse peritoneum is shown in figure 3. The data represent an average of four experiments, with three animals for each period (except for results on the fate of phagocytized bacteria, which are derived from one of the experiments). When the whole suspension was considered, counts of the parental form increased progressively, while counts of the revertant form decreased with time. When the fate of phagocytized bacteria was examined alone, both parental and revertant forms increased in number for the first 2 hr. The parental form then continued to increase in number, while the number of viable revertant bacteria declined markedly. Surface phagocytosis. No significant difference between the proportions of leukocytes with ingested bacteria or in numbers of bacteria ingested per leukocyte was noted for revertant and parent bacteria. With both PTI09 and RLTI09, logarithmic growth-phase bacteria were more easily phagocytized than organisms in the stationary phase. Tube phagocytosis. Logarithmic-phase organisms were used in an experiment designed for the study of both phagocytosis and the fate of the ingested organisms. The results (average of three experiments) are shown in figure 4. The parent
Downloaded from http://jid.oxfordjournals.org/ at University of Birmingham on August 31, 2015
forms. The biochemical tests that had positive results were: Kligler iron agar; Voges-Proskauer; nitrite reduction; and fermentation of mannitol, xylose, arabinose, raffinose, sorbitol, sucrose, glucose, lactose, mannose, galactose, maltose, trehalose, salicin, and levulose. Oxidase, indole, and ornithine reactions were negative. Inulin, dulcitol, and glycerol were not fermented. Litmus milk was acidified but not coagulated. Gelatin was not liquefied, H 2 S was not produced, and urease activity was not demonstrated. Growth of the revertant in BHI broth (figure 2) was distinctly slower than that of the parent during the early phase, although both forms reached the same level by 18 hr. Difference in the rate of growth was even more striking when the organisms were tested in Hanks' solution. The parental form grew somewhat more slowly in Hanks' solution than in BHI broth and reached a level of only 107 Iml in Hanks' solution as compared with 109 /ml in BHI broth. However, the revertant did not grow at all in Hanks' solution. Size of bacteria. The average of the transverse diameters of 20 individual PT109 organisms was 4.1 11m, including the capsule; for RLT109 organisms this value was 1.2 11m. Both types of bacterium without capsules had diameters of 1.0 11 m . Weights of cell envelopes were as follows: 2.3 X 1010 parent bacteria yielded 1.95 mg; 2.7 X 1010 revertant bacteria yielded 0.61 mg. This finding indicated that the cell envelope was about 3.5 times larger in the parental form than in the revertant.
Guze, Harwick, and Kalmanson
249
Klebsiella L-Form Revertant Virulence
400
~
ffi
300
c
I-
~
ll..
o
...J
~ 200
~ ::l In
I-
Z
W U
a:::
~ 100
Discussion
0 ......-e5=t=:==------r---.-----r2 3 4 TIME
HOURS
Figure 3. Phagocytosis and survival of parent and revertant Klebsiella pneumoniae in mouse peritoneum. (0--0) = revertant in whole suspension; (0- - -0) = parent in whole suspension; (0--0) = phagocytized parent organisms only; (0- - -0) = phagocytized revertant organisms only; (e--e) = percentage of parent organisms phagocytized; (e- - -e) = percentage of revertant organisms phagocytized.
Prolonged cultivation as a penicillin-induced Lform and periodic reversion to the bacterial form resulted in a step-by-step reduction in virulence of an originally virulent strain of K. pneumoniae in the mouse. After 109 L-form passages, virulence was totally lost and was not restored by 40 passages through mouse peritoneum. This loss
,
I
3000 f
oct
b «
ilID
f
:E: ~
"
~
I
ii:
~ 2500
,
CD
grew vigorously in the diluent without any leuko~ cytes present, while the revertant did not grow. The viable count of parent bacteria in the mixture increased more rapidly than did the count in the control culture, as if some growth-stimulating effect occurred as a result of the presence of leukocytes. On the other hand, the count of viable revertants did not change with time. The parental form obviously grew after phagocytosis, while the revertant did not. At each interval there was more phagocytosis of PTI09 than of RLTI09. Thus the parental form was phagocytized more efficiently than the revertant, but, once engulfed, the parent actually grew within the leukocyte, while the revertant merely survived during the period studied. Bactericidal susceptibility. There was no evidence of killing of either logarithmic- or stationary-phase parent or revertant forms by human or mouse serum after incubation for 1 hr.
~ !AI
f
,f
f
~ 2000
I
...J
~
I
I
f
~ l-
>-
v
... ~ ... e
/ ... / ' " ,.....
~ 1500 :::>
,e/,'/ ,'/ "" /'
~
/'
/,,,
~
500
"""
e ;'
",/' .../.d'
o
I
/'
60
d ::l
~
40 ~
~
/
_---_-D- _------D
100.pf1ftff-;:----
~ ~
...",,,,/ ",.e~ ;f ~
~ 1000
8
::l II::
~I
en
80
_-«/
~
~
2
3
~
20
Z
I-
~
TIME - HOURS
Figure 4. Phagocytosis and survival of parent and revertant Klebsiella pneumoniae in vitro. (x--x) = revertant in control medium (diluent without leukocytes); (x- - -x) = parent in control medium; (0--0) revertant in whole suspension; (0- - -0) parent in whole suspension; (0--0) = phagocytized revertant organisms only; (0- - -0) = phagocytized parent organisms only; (e--e) = percentage of revertant organisms phagocytized; (e- - -e) = percentage of parent organisms phagocytized.
Downloaded from http://jid.oxfordjournals.org/ at University of Birmingham on August 31, 2015
Immunization with RLT109 and challenge with PT109. The LD50 was < 10 bacteria in unimmunized controls. In contrast, the LD50 in immunized mice was 1.08 X 105 bacteria, a value indicating a high degree of protection (figure 5). Immunization with RLT109 and challenge with heterologous strain no. 276. The outcome of inoculation of RLTI09 is shown in figure 6. The LD50 was 170 organisms in unimmunized control mice, 610 organisms in those immunized with heat-killed RLTI09, and 40 organisms in those immunized with live RLT 109. These differences were not statistically significant.
Guze, Harwick, and Kalmanson
250
6.00
au
IMMUNIZED WITH
g
REVERTED.
(/)
-.
LIVE
..J
