Vol. 130, No. 2

JOURNAL OF BACTERIOLOGY, May 1977, p. 939-942

Printed in U.S.A.

Copyright C© 1977 American Society for Microbiology

Stepwise Selection of Defective Nitrogen-Fixing Phenotypes in Escherichia coli K-12 by Dimethyl Sulfoxide MARY L. SKOTNICKI AND BARRY G. ROLFE* Department of Genetics, Research School of Biological Sciences, The Australian National University, Canberra City, A.C.T. 2601, Australia

Received for publication 10 August 1976

Storage in dimethyl sulfoxide of Escherichia coli K-12 hybrids carrying Klebsiella pneumoniae nif+ genes can result in selection of a defective nitrogenfixing phenotype. Dimethyl sulfoxide appears to affect particular inner membrane proteins associated with energy metabolism in E. coli K-12. Nitrogen fixation in nature is classically a process in which atmospheric nitrogen is converted to ammonia by the enzyme nitrogenase (specified by nif genes) (19). The presence of nitrogenase in bacteria is commonly measured by their growth on nitrogen-free medium and by reduction of acetylene to ethylene (18). Escherichia coli is phenotypically a non-nitrogen fixer and is generally regarded as being genetically deleted of nif functions (21). However, various hybrids between Klebsiella pneumoniae strain M5al (a nitrogen fixer) and E. coli C and E. coli K-12 have been made and shown to fix nitrogen, indicating successful transfer and expression of nifKp genes (nif genes from K. pneumoniae) in an E. coli background (6, 9). To study the influence that various mutations in energy metabolism have on the expression of nifKp genes in E. coli K-12, a variety of hybrid strains was constructed in respiratory mutants uncoupled in oxidative phosphorylation (unc) (7) and mutants defective in anaerobic electron transport (nitrate reductase system, chlA-G) (3, 20) (Table 1). This communication reports the effect of storing these hybrids in dimethyl sulfoxide (Me2SO) and the subsequent expression of nitrogen fixation in E. coli K-12. Me2SO is widely used as a cryoprotective agent (2), but it has recently been shown to affect recombination (12), mutagenesis (1), deoxyribonucleic acid denaturation (16), the initiation of specific ribonucleic acid transcripts (17), and specific membrane functions (13). Hybrid strains freshly grown in Luria broth (LB) containing 0.5% glucose (14) were stored in the same broth containing 7% Me2SO (Sigma grade 1, lot 102C-2570) at -20°C. Samples of these frozen cultures were regularly checked over a period of 15 months. Cultures were diluted into fresh broth and regrown when required (15). These cultures of hybrid strains

were then tested for their original properties and for the presence of the FN68 plasmid by resistance to carbenicillin and by their ability to grow on solid nitrogen-free medium (NFM) (2) and in liquid NFM (in Pankhurst tubes) and ability to reduce acetylene to ethylene. As a control to test for the presence of plasmid-containing cells, samples of Pankhurst tube cultures were routinely checked by plating on solid NFM and LB glucose plates. Colony numbers were similar on both media, and colony sizes on NFM always reflected the particular phenotypic stage of hybrids. Occasionally, colonies from these NFM plates were restreaked onto fresh NFM plates, and regrowth was always of the same extent as observed on the NFM plate from which the inoculum was isolated. Well-isolated single colonies from both the NFM and LB glucose plates were tested for the presence of the FN68 plasmid by resuspending colonies in buffer and directly checking their drug resistance and nitrogen-fixing properties. These tests eliminated any possibility of plasmid loss or cross-feeding as an explanation of the Me2SO effect. On several occasions frozen cultures were directly plated onto NFM and LB glucose plates. Again, well-isolated single colonies were directly tested for their nitrogen-fixing capacities. Changes in phenotypic expression observed by this method were similar to those found for cultures grown in broth before testing. The effect of Me2SO on stored hybrid strains is shown in Fig. la. The responses of the strains can be divided into two classes: class I hybrid strains, which stably express all their properties at the same level as when they were first stored, and class II hybrids, which progressively lose their nitrogen-fixing capacity even though the carbr plasmid is still present and can be mobilized into other E. coli K-12 recipi-

939

940

J. B3ACTERIOL.

NOTES

TABLE 1. Effect of Me2SO on expression of nif+KP genes and viability of hybridsa Class of Bacterium

Hybrid strain

Relevant mutation

nif*Kp expression"

Viabilityc

after storage in Me2SO

7 months

or:

15 months

II 10-4 AN249 (FN68) uncA10-6 I 5 x 10-2 AN283 (FN68) uncB10-4 AN285 (FN68) unc-405 I NrT' 10-4 C181 (FN68) chIAII 10-510-7_1s-8 C183 (FN68) II chMBNT 10-7 C113 (FN68) chlDI 5 x 10-3 10-l C123 (FN68) II chlDNT 10-7 II 10-5 DD27 (FN68) chlDA 10-7_10-8 II DD182 (FN68) NT chlAchlDA 10-7 C197 (FN68) chMMII NT 10-7_10-8 I NT 4 x 10-3 DD38 (FN68) chlGII 10-4 RB101 chlAchlDA 10-6 E. coli C C-M7 I NT 5 x 10-3 K. pneumoniae M5al I 4 x 10-l 10-2 a E. coli K-12 strain SB1801, containing the FN68 plasmid (F'nif+Kp, his+, carbr), was used as the donor in Conjugation experiments (4), because in addition to the nifKp genes this plasmid contains a stably linked carbenicillin resistance marker (carbr) (8). Hybrids were constructed by selecting for transfer of carbenicillin resistance and then testing for cotransfer of nif+Kp genes by growth on NFM and acetylene reduction. In all clones tested, the carbr marker did not segregate from the nifrKP genes. Where necessary, NFM medium was supplemented with appropriate amino acids at 25 ,ug/ml and vitamins at 10-3 ,uM. b Class I of nif+Kp expression is recognizable by stable expression of these genes; class II hybrids show stepwise decline in nif+Kp gene expression (see Fig. 1). c Viability at 7 and 15 months is expressed as a fraction of the original stored cells and was measured on LB glucose and NFM plates. d NT, Not tested.

E. coli K-12

ent cells (previously stored in glycerol) to give Nif+ isolates with complete expression. The various class II hybrids carrying the FN68 plasmid show a similar pattern of loss of the Nif+ phenotype, although the onset of this decline varies with the hybrid strain and ranges from 1 to 6 months. The decline in these hybrids is first characterized by good growth on solid and liquid NFM but complete inability to reduce acetylene to ethylene. This step is closely followed by a drop of about 60% in the extent of growth on solid and liquid NFM. The final step in the evolution of a defective nitrogen-fixing phenotype is characterized by very poor or no growth in liquid NFM and formation of small colonies on solid NFM. Class I and II responses were found with both unc-(nifrKp) and chl-(nifrKp) hybrid strains. The effect is: (i) specific to Me2SO storage, as the same hybrid strains stored in glycerol (14) over a similar length of time showed no loss of nitrogen-fixing capacity; (ii) apparently specific to some E. coli K-12 nif*Kp hybrids but not limited to any one particular E. coli K-12 background, as the loss of nitrogen-fixing capacity occurs in a variety of E. coli K-12 nif+Kp hybrids (Table 1); (iii) not limited only to E. coli K-12 hybrids, which carry the nif+KP genes on a plasmid, since a modified class II-type response occurred when the nifrKp genes were integrated

into the chromosome, as in transductant strain RB101 (Fig. lb). In contrast to the plasmidbearing hybrids, RB101 does not show the "allor-none" response for acetylene reduction, but rather a gradual decline in both acetylene reduction and growth in liquid NFM. Good growth on solid NFM remains for a longer period, but after 15 months of storage in Me2SO strain RB101 has a completely Nif- phenotype. However, since the nifrKp genes could be transduced from the Nif- RB101 back into the original recipient host KS649, with expression, Me2SO does not directly affect the transferred nif+Kp operon in the background of E. coli K-12. Over the 15-month storage period the chromosomally integrated E. coli hybrid C-M7 and K. pneumoniae strain M5al were stable with respect to their nitrogen-fixing properties. When the Me2SO-stored hybrids were assayed for viability after 15 months, class I strains still had a viability of 10-2 to 10-4 of original stored cells, whereas the survival level was 10-6 to 10-8 for class II strains. Although the surviving cells of class II hybrids were still carbenicillin resistant and formed small colonies on solid NFM, they were unable to grow in liquid NFM and reduce acetylene. Since the cultures grown up from these survivors still have the FN68 plasmid, the simplest explanation is that storage in Me2SO

VOL. 130, 1977

NOTES

941

50 40

30

c--4 a, ~ 20 a:

E

10

U-

z

=

z n

C,

0 0

50

_,

O

0

0

0

C,

u

30} O

20

.................i.....

L6 Time

0

9

(months)

FIG. 1. Expression of nif+KP genes in E. coli K-12 hybrid strains over a period of 15 months of storage in Mej,0 was tested by: growth on solid NFM under 99% N211 % CO2 (11) after 7 days (@ 0); growth in liquid NFM in Pankhurst tubes (5, 18) (---- 0); and acetylene reduction to ethylene (23), both with and without 100 pg of vitamin-free Casamino Acids per ml ( . 0). All incubations were at 30°C. (a) Maintenance of expression of nitKP genes on plasmid FN68 in E. coli K-12 hybrids falls into two classes: class I (open circles): stable expression of growth on solid and in liquid NFM and of acetylene reduction [for example, the uncB- hybrid AN283(FN68)]; class II (closed circles): gradual decline of expression of nifrK genes shown first by loss ofacetylene reduction and then by poor or no growth in liquid NFM, resulting finally in hybrids that can only form small colonies on solid NFM [example shown is the chlA - strain C181 (FN68)]. K. pneumoniae M5al and E. coli C hybrid C-M7, with nif+K, genes stably integrated into the his region of the chromosome, are class I strains. (b) Decline of expression of nifrK, genes in transductant strain RB1 01, an E. coli K-12 hybrid between C-M7 and E. coli K-12 strain KS649 (22), with the his nifK, region stably integrated into the chromosome. This strain was constructed by P1 transduction (14) from host C-M7. At 15 months, the Nif phenotype could be restored by transducing the his nif region of the chromosome back into the original recipient KS649.

causes killing of the majority of hybrid cells with the complete Nif+ phenotype, leaving only cells that have a defective nitrogen-fixing phenotype. One test of this hypothesis is to take a parental strain stored in Me2SO for 15 months and construct a nif+Kp hybrid similar to the equivalent one that had been stored in Me2SO over the same period and then compare the nitrogenfixing properties of the two hybrids (Table 2). Parental strains C181 and C181-1 (the same strain stored in Me2SO for 15 months) do not fix nitrogen. The original hybrid C181(FN68) had a complete nitrogen-fixing phenotype when first made, but gave a class II response on

storage in Me2SO, growing poorly on solid and liquid NFM. The newly constructed hybrid C181-1(FN68), however, never reduced acetylene and could only grow sparingly on solid and liquid NFM. Thus, storage of E. coli K-12 strains in Me2SO can select for cells that are defective in their expression of nif+Kp genes, even before these genes are transferred (Table 2). However, the uncB- and unc-405 hybrids, which are Me2SO resistant, show no loss of the Nif+ phenotype. Both uncB and unc-405 mutations are alterations within the inner membrane of E. coli K-12, which contrasts with the uncA - mutation (defective in membrane-bound

942

J. BACTZCRIOL.

NOTES

TABLE 2. Effect of Me,SO on nifrKp expression by E. coli K-12 for 15 monthsa

C2H2 re-

Growth on on solid solid duction in (nmol of in liquid NFM unNFM der 99% C2HS/ omPeter CO2 (colminlmg (nephelN2I1% min/mg 0 Cl (months) of pro- oee size units) ony tein) in mm) 0 0 0 0 C181 32 2.0 0 45 C181 (FN68) 4 0.5 15 0 C181 (FN68) 0 0 0 C181-1 15 0.5 0 0 8 C181-1 (FN68) growth on solid and liquid aAcetylene reduction and NFM was measured as described in Fig. 1. C181, which had been stored in Me%SO for 15 months (C181-1), was used as a recipient strain with donor strain SB1801 (FN68) in a conjugation experiment similar to that for C181 before Me2SO storage. C181-1 (FN68) was tested for expression of nif/KP genes.

Length of storStrain age in StraingeessO Me2SO

[Mg2+-Ca2+] adenosine triphosphatase), whose background is isogenic with that of the uncBand unc-405 strains (7) but whose hybrid showed a class II response (Table 1). This result, and the stability of some chlorate-resistant hybrids, implies that the Me2SO effect is on particular inner membrane proteins associated with energy metabolism inE. coli K-12 and is a potential probe of the energy-coupling site needed for nif+ expression in E. coli K-12. In liquid NFM with 5 mM (NH4)2S04, all hybrid strains grew well over the 15-month period, which shows that Me2SO induces a specific effect on nif expression and not an indirect one. Development of the Nif- phenotype suggests a multistep process giving a gradual loss of expression of the nif+Kp operon in E. coli K-12, similar to evolution of penicillin resistance in bacteria by cell envelope alterations (10). Recognition of this sequential pattern has enabled the isolation of various E. coli K-12 mutants that exhibit this hierarchy of expression of nitrogen-fixing capacity. M.L.S. is the recipient of a Commonwealth Scholarship and Fellowship Plan Award and a Ph.D. Scholarship, Australian National University. The following people are thanked for cultures: G. B. Cox, R. A. Dixon, D. Dykhuizen, F. Gibson, M. Gottesman, and W. A. Venables. We thank P. M. Gresshoff and K. Williams for helpful criticisms.

LITERATURE CITED 1. Anwar, S. Y., and G. M. Raddy. 1975. Effect ofdimethyl sulphoxide in combination treatments with diethyl sulphate in the induction of chlorophyll mutations in Oryz sativa L. Indian J. Exp. Biol. 13:187-188. 2. Ashwood-Smith, M. J. 1967. Radioprotective and cryoprotective properties of dimethyl sulfoxide in cellular systems. Ann. N.Y. Acad. Sci. 141:45-62. 3. Bachmann, B. J., K. B. Low, and A. L. Taylor. 1976.

Recalibrated linkage map of Escherichia coli K-12.

Bacteriol. Rev. 40:116-167. 4. Bernstein, A., B. G. Rolfe, and K. Onodera. 1972. Pleiotropic properties and genetic organization of the toLA,B locus of Escherichia coli K-12. J. Bacteriol. 112:74-83. 5. Campbell, N. E. R., and H. J. Evans. 1969. Use of Pankhurst tubes to assay acetylene reduction by facultative and anaerobic nitrogen-fixing bacteria. Can. J. Microbiol. 15:1342-1343. 6. Cannon, F. C., R. A. Dixon, J. R. Postgate, and S. B. Primrose. 1974. Chromosomal integration of Klebsiella nitrogen fixation genes in Eacherichia coli. J. Gen. Microbiol. 80:227-239. 7. Cox, G. B., and F. Gibson. 1974. Studies on electron transport and energy-linked reactions using mutants of Escherichia coli. Biochim. Biophys. Acta 346:1-25. 8. Dixon, R. A., F. C. Cannon, and A. Kondorosi. 1975. Derivation of a P-type R factor carrying the nitrogen fixation genes from Klebsiella pneumoniae: nif expresaion in unrelated bacteria, p. 43. In Proceedings of the Society of General Microbiology, vol. 2. 9. Dixon, R. A., and J. R. Postgate. 1972. Genetic transfer of nitrogen fixation from Klebsiella pneumoniae to Escherichia coli. Nature (London) 237:102-103. 10. Erikason-Grennberg, K. G. 1968. Resistance of Escherichia coli to penicillins. II. An improved mapping of the ampA gene. Genet. Res. 12:147-156. 11. Hill, S. 1973. A simple method for exposing bacterial cultures on solid media to a defined gas mixture using nylon bags. Lab. Pract. 22:193. 12. Ihrke, C. A., and W. E. Kronstad. 1975. Genetic recombination in maize as affected by ethylenediaminetetracetic acid and dimethyl sulfoxide. Crop Sci. 15:429431. 13. Kunze, M. 1974. Fermentation of lactose by permeaseless E. coli in the presence of dimethylsulfoxide. Zen-

tralbl. Bakteriol. Parasitenkd. Infektionskr. Hyg. Abt. 1 Orig. Reihe A 229:352-355. 14. Miller, J. H. 1972. Experiments in molecular genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 15. Monk, M., M. Peacey, and J. D. Gross. 1971. Repair of damage induced by ultraviolet light in DNA polymerase-defective Eacherichia coli cells. J. Mol. Biol.

58:6234630.

16. Nakanishi, S., S. Adhya, M. Gottesman, and I. Psatan. 1974. Activation of tranacription at specific promotors by glycerol. J. Biol. Chem. 249:4060-4066. 17. Nakanlshi, S., S. Adhya, M. Gottesman, and I. Patan. 1974. Effects of dimethylsulfoxide on the E. coli gal operon and on bacteriophage lambda in vivo. Cell 3:39-46. 18. Postgate, J. R. 1972. The acetylene reduction test for nitrogen fixation, p. 343-356. In J. R. Norris and D. W. Ribbons (ed.), Methods in microbiology, vol. 6B. Academic Press Inc., London. 19. Postgate, J. R. 1974. Prerequisites for biological nitrogen fixation in free-living heterotrophic bacteria, p. 663-686. In A. Quispel (ed.), Biological nitrogen fixation. North Holland, Amsterdam. 20. Rolfe, B. G., and K. Onodera. 1972. Genes, enzymes and membrane proteins of the nitrate respiration system ofEscherichia coli. J. Membr. Biol. 9:196-207. 21. Shanmugam, K. T., and R. C. Valentine. 1975. Molecular biology of nitrogen fixation. Science 187:919-924. 22. Shimnda, K, R. A. Weisberg, and M. E. Gotterman. 1972. Prophage lambda at unusual chromosomal locations. I. Location of the secondary attachment sites and the properties of the lysogens. J. Mol. Biol. 63:483-603. 23. Tubb, R. S., and J. R. Postgate. 1973. Control of nitrogenase synthesis in Kiebsiella pneumoniae. J. Gen. Microbiol. 79:103-117.