Isolation and complementation of mutants of Anabaena sp. strain PCC ...

JOURNAL

OF

Vol. 170, No. 3

BACTERIOLOGY, Mar. 1988, p. 1239-1244

0021-9193/88/031239-06$02.00/0 Copyright C 1988, American Society for Microbiology

Isolation and Complementation of Mutants of Anabaena sp. Strain PCC 7120 Unable to Grow Aerobically on Dinitrogen C. PETER WOLK,1* YUPING CAI,1 LILIANA CARDEMIL,1t ENRIQUE FLORES,'t BARBARA HOHN,2 MARCIA MURRY,1 GEORG SCHMETTERER,1§ BERNHARD SCHRAUTEMEIER,1 AND RUTH WILSON1 MS U-DOE Plant Research Laboratory, Michigan State University, East Lansing, Michigan 48824,1 and Friedrich Miescher-Institut, CH-4002 Basel, Switzerland2 Received 2 October 1987/Accepted 14 December 1987

Mutants of Anabaena sp. strain PCC 7120 unable to grow aerobically on dinitrogen were isolated by mutagenesis with UV irradiation, followed by a period of incubation in yellow light and then by penicillin enrichment. A cosmid vector, pRL25C, containing replicons functional in Escherichia coli and in Anabaena species was constructed. DNA from wild-type Anabaena sp. strain PCC 7120 was partially digested with Sau3AI, and size-fractionated fragments about 40 kilobases (kb) in length were ligated into the phosphatasetreated unique BamHI site of pRL25C. A library of 1,054 cosmid clones was generated in E. coli DH1 bearing helper plasmid pDS4101. A derivative of conjugative plasmid RP-4 was transferred to this library by conjugation, and the library was replicated to lawns of mutant Anabaena strains with defects in the polysaccharide layer of the envelopes of the heterocysts. Mutant EF116 was complemented by five cosmids, three of which were subjected to detailed restriction mapping; a 2.8-kb fragment of DNA derived from one of the cosmids was found to complement EF116. Mutant EF113 was complemented by a single cosmid, which was also restriction mapped, and was shown to be complemented by a 4.8-kb fragment of DNA derived from this cosmid.

Under aerobic conditions in the absence of fixed nitrogen, a low percentage of the vegetative cells of certain filamentous cyanobacteria differentiate into cells, called heterocysts, that have a thick envelope.The envelope contains, outside of the peptidoglycan layer, a laminated layer of glycolipids and an outer, homogeneous layer of polysaccharide (26). The ability of these organisms to grow on N2 is attributable to nitrogen fixation taking place within the heterocysts (18). The envelope is thought to reduce the rate at which oxygen enters the heterocysts so extensively that the oxidases present in those cells can reduce the oxygen that enters (26). Indeed, mutants lacking the envelope glycolipids (10) or having an irregularly deposited polysaccharide layer (4) were unable to grow aerobically on N2, and N2-fixing revertants had regained morphologically normal envelope layers. The first indication at an ultrastructural level that a vegetative cell is differentiating into a heterocyst is the appearance at the periphery of the cell of a fibrous layer (15, 24) that may be chemically indistinguishable from the polysaccharide that forms the homogeneous layer of the envelope. In the three species studied, this polysaccharide has a long backbone, consisting of repetitions of the 1,3-

biochemically but were not amenable to genetic analysis. The finding that RP-4-mediated conjugative transfer of pBR322-based plasmids takes place from Escherichia coli to Anabaena species in the presence of helper plasmids (derivatives of ColD or ColK [28]) provided the opportunity to identify genes mutated in Anabaena spp. by complementation of mutations with cloned DNA. We illustrate this possibility by using two mutants that produce heterocysts having irregular envelope polysaccharide layers. MATERIALS AND METHODS Isolation of mutants. Anabaena sp. strain PCC 7120 was grown in medium AA/8 with 5 mM nitrate (13) to a density of 4 ,ug of chlorophyll (Chl) ml-'. A culture was sedimented and then resuspended at 2 ,ug of Chl ml-' in fresh AA/8 nitrate medium. Subsequent operations were performed with illumination from yellow lamps (14). A 7.5-ml portion of the suspension of cells was irradiated in a 90-mm-diameter petri dish for 10 min at a distance of 45 cm from a 30-W germicidal (UV) lamp (General Electric Co., Schenectady, N.Y.) that had been prewarmed for 0.5 h. A 1.5-ml subsample was incubated for 36 h at 30°C in yellow light. As determined by growth of serially diluted samples on agar in white light, the viability of the UV-treated filaments had been reduced by a factor of 5 x 103. One milliliter of the 1.5-ml subsample was diluted with 50 ml of AA/8 nitrate and grown in white light for ca. 15 generations to permit segregation and expression of mutations. The resulting culture was cavitated in a sonic cleaning bath for 30 s, giving filaments with an average of 2.9 cells per filament. The fragmented filaments were suspended (0.25 ,ug of Chl ml-1) in 50 ml of AA/8 medium and incubated under growth conditions for 48 h, giving a culture with 0.54 ,ug of Chl ml-'. Penicillin G was added to 200 U ml-'. After 72 h of incubation under growth conditions, the culture appeared lysed. Surviving cells were harvested, washed with

linked tetrasaccharide mannosyl-glucosyl-glucosyl-glucose, and has many short side branches (1-3). Until recently, mutants of heterocyst-forming cyanobacteria (4, 8, 10, 23, 25) could be analyzed physiologically and * Corresponding author. t Present address: Departamento de Biologia, Facultad de Ciencias, Casilla 653, Universidad de Chile, Santiago, Chile. t Present address: Departamento de Bioquimica, Facultad de Biologia y Consejo Superior de Investigaciones Cientificas, E-41080 Sevilla, Spain. § Present address: Institute of Physical Chemistry, University of Vienna, A-1090 Vienna, Austria.

1239

1240

WOLK ET AL.

FIG. 1. Cosmid shuttle vector pRL25C. Only restriction sites that are unique or are potentially useful for cloning are shown.

AA/8 medium, and plated with soft (0.5%) agar atop AA nitrate 1% agar. A total of 100 of the ca. 400 colonies that grew were transferred on toothpicks to BG-11. and BG-11 solid media (20). Twenty-three colonies (strains EF101 through EF123) showing no growth on medium lacking nitrate were streaked to, and grown on, BG-11 solid medium. Material preserved with 1.85% formaldehyde was photographed with a Zeiss photomicroscope under oil immersion. Methods for analysis of lipids (17) and for measurement of reversion frequencies (4) have been described; nitrogenase activity was measured (10) under 10% C2H2-90% Ar in the presence of 20 FM DCMU [3-(3,4-dichlorophenyl)-1,1-

dimethyl urea]. Construction of cosmid shuttle vector pRL25C. Plasmid pRL7, constructed identically to pRL6 (28) but with the neomycin phosphotransferase (neo) gene inserted into pRL1 in the opposite orientation, has MstI sites to both sides of the chloramphenicol acetyl transferase gene (cat) plus a third within the neo gene. Religation of pRL7 partially restricted with MstI (New England BioLabs, Beverly, Mass.), and screening of transformants produced Nmr Cms plasmid pRL7A. All four XhoII sites of pRL1 and of its noncyanobacterial portion, pRLlV (28), are clustered in and near its E. coli oriV region. Screening of transformants from the products of pRLlV that had been partially digested with XhoII and then religated identified a derivative, pRLlVB, for which deletion between base pair (bp) 3,213 and 3,225 of the parental pBR322 had given rise to a unique BamHI site (the original BamHI site had been destroyed during the construction of pRL1). Substitution of the 1,274-bp NdeIEcoRI fragment from pRLlVB containing this BamHI site for the corresponding sequence of pRL7A gave rise to a shuttle vector, pRL25, with a BamHI site for cloning. A 403-bp cos-containing fragment was excised from plasmid pWH5 (constructed together with pWH4 [11] but with the cos site in the opposite orientation) with AccI and was inserted into the unique ClaI site of pRL25, giving cosmid shuttle vector pRL25C (Fig. 1).

J. BACTERIOL.

Formation of a cosmid library of Anabaena sp. strain PCC 7120. Chromosomal DNA of high molecular weight was isolated from nitrate-grown Anabaena sp. strain PCC 7120, partially restricted with Sau3AI, sized on gradients of NaCl, and ligated to BamHI-cut, phosphatase-treated pRL25C as previously described (12). DNA packaged in vitro (16) was plated on E. coli DH1 bearing helper plasmid pDS4101 (ColK: :Tnl) with selection for Kmr Apr. The resulting colonies were transferred on toothpicks to wells of microtiter plates, containing 100 pl of L broth with 50 jig ml-' each of kanamycin and ampicillin, and were grown overnight. Each well was then supplemented with 120 ,lI of 80% (vol) glycerol. The 11 microtiter plates constituting the library (ca. 6.5 genome equivalents) were stored at -75°C. Mating of bacteria with Anabaena sp. strain PCC 7120. The cells, still frozen, were replicated to microtiter plates containing, per well, 175 ,ul of L broth with 50 jig ml-' each of kanamycin and ampicillin. After overnight growth at 37°C, each well was supplemented with 50 pAl of an exponentialphase culture of E. coli DH1 bearing a conjugative plasmid, permitting unselected transfer to the cosmid-bearing strains. We used, as conjugative plasmid, RP-4 or a Kms derivative of it, pRL443 (J. Elhai and C. P. Wolk, unpublished observations). A 200-ml, actively growing but light-limited culture of mutant EF116 or EF113 in liquid medium AA/8 nitrate was filtered onto a prewetted sheet of nitrocellulose (23 by 23 cm; HAWP 000 10; Millipore Corp., Bedford, Mass.). A pronging device was then used to transfer 2-pul portions of the mated bacteria to the lawn of Anabaena sp. strain PCC 7120. The sheet of nitrocellulose was transferred first to AA nitrate agar supplemented with 5% L agar (6 h), then to AA nitrate agar (18 h), and finally to AA agar containing 25 ,ug of neomycin ml-', all in tissue culture plates (Nunc, Roskilde, Denmark) and were incubated in the light at 30°C. Presumptive exconjugants were freed from E. coli by streaking on selective medium. Cosmids were isolated from exconjugant cyanobacteria by the method of Simon (22). Mapping and subcloning of complementing cosmids. Cosmids from library elements identified as complementing particular mutants were repackaged, plated on E. coli HB101 free of other plasmids, and subjected to conventional restriction mapping. We find it convenient to screen for enzymes that have few sites in the insert DNA and to map these sites. To subclone, partial Sau3AI digests of complementary cosmids were sized on 0.6% gels of GTG agarose (FMC BioProducts, Rockland, Maine) to ca. 3 to 6 kilobases (kb), eluted with DEAE-cellulose (NA45:Schleicher & Schuell, Inc., Keene, N.H.), and ligated to BamHI-cut, phosphatasetreated pRL25 or pRL25C.

RESULTS Strains EF101 through EF123 proved able to grow on medium BG-11 but not on medium BG-110. These 23 strains incapable of aerobic growth on N2 represented at least nine mutant phenotypes. In EF116, the heterocysts formed an outer layer of polysaccharide which differed from the corresponding layer in heterocysts of the wild-type strain in that it was less cohesive (Fig. 2b). In EF113, portions of the polysaccharide layer were observed only near the ends of the heterocysts. EF114 and EF122 lacked the major glycolipid characteristic of the laminated layer of the heterocyst envelope (Fig. 2c and 3). Few heterocysts of EF111 remained attached when filaments were starved for fixed nitrogen. Finally, EF104 and EF121 were not morphologically distinguishable from the wild-type strain. Other char-

VOL. 170, 1988

acteristics of mutants representative of the different phenotypes are presented in Table 1. The reversion data suggest that the observed phenotypes were the result of single mutations. Complementation studies were carried out with strains EF113 and EF116 because of their low reversion frequencies, because of our long-standing interest in the heterocyst envelope (1-3), and because of their potential utility for developmental studies (see Discussion). Selection for complementing cosmids could not be simply for N2 fixation because nitrogen present from decaying bacteria and cyanobacteria allowed growth. Neither was it sufficient to select only for Nmr, because many colonies, having received an Nmr shuttle vector, grew. However, the dual selection yielded a very clear result (Fig. 4). Five cosmids (41B9, 41C1, 41E11, 43E7, and 43G9) complemented mutant EF116, and a single cosmid (43G5) complemented mutant EF113. The complementing cosmids could be recovered by freeing the cyanobacteria of E. coli, isolating plasmid, packaging, and plating on E. coli, as well as from the original library. Restriction maps of the overlapping inserts of three of the cosmids complementing EF116 are presented in Fig. 5, and a restriction map of 43G5 is presented in Fig. 6. A subclone (pRW6) of a 4.8-kb fragment of 43G5 that still complements EF113 contains the sole EcoRI site of the insert and has no site for BamHI. Two subclones (subclones 11 and 17) of 41B9 that also complemented EF116 contained the unique NruI site common to 41B9, 43E7, and 43E11. Plasmid 41B9

MUTANTS OF ANABAENA SP. STRAIN PCC 7120

I.1-11-1--.

1241

,-,qmp-qwlw Awzm*,-.

GL

Or a b FIG. 3. Chromatogram of lipids extracted from wild-type Anabaena sp. strain PCC 7120 (b) and mutant EF114 (a). The major glycolipid (GL) derived from the laminated layer of the heterocyst envelope is indicated. Or, Origin.

subclone 17 had regenerated the BamHI site of the vector at one end of the insert, apparently lacked internal DraI sites, and had a DraI site 7 bp beyond the other end of the insert. The 2.8-kb BamHI-DraI fragment, subcloned to positive selection shuttle vector pRL57 (J. Elhai and C. P. Wolk, submitted for publication), still complemented. Insertion of a fragment of DNA containing a streptomycin resistance determinant (5) in either orientation within the unique NruI site in subclone 17 (yielding plasmids pRL61La and -b) prevented general complementation, but colonies formed that contained both yellow and green regions, suggestive of recombination having taken place.

FIG. 2. Anabaena sp. strain PCC 7120 wild type (a) and mutants EF116 (b) and EF114 (c), showing structural differences in the envelope layers of the heterocysts (H). The outer, polysaccharide layer of the envelope of the heterocyst of EF116 is less cohesive than is the corresponding layer in the wild-type strain; a gap beneath that layer in EF114 correlates with a deficiency of glycolipids (see also Fig. 3). Magnification, x 1,620 for panels a and b and x 1,000 for panel c.

DISCUSSION We have isolated a number of mutants from UV-irradiated Anabaena sp. strain PCC 7120. UV-irradiated microorganisms are normally grown in the dark to permit mutational repair of damaged DNA before they are exposed to photoreactivating white light (9). Obligately photoautotrophic cyanobacteria, such as Anabaena sp. strain PCC 7120, cannot be grown in the dark. However, only light of wavelength of 10-6 2 x 10-7 >10-6 4 x 10-8 3 x 10-7 4 x 10-8 >10-6 4 x 10-7

No difference observed Heterocyst envelope often deformed Heterocyst envelope very nonevenly deposited Attached heterocysts rare Heterocyst has only vestigial envelope Gap in heterocyst envelope (lacks major glycolipid) Heterocyst envelope polysaccharide less cohesive than normal Heterocyst contents unusually pale and granular No difference observed

+ + tr + + + + +

M. Murry and C. P. Wolk, manuscript in preparation.

the mutation in EF117, mutations more commonly affected the heterocyst envelope, that is, the cellular organelle that presumably serves as the principal barrier to the movement of 02 to the sites of nitrogen fixation. The envelope largely failed to form (EF113), was deposited irregularly (EF107 and EF110), or had a defect that appeared specific to the glycolipid layer (EF114 and EF122) or to the polysaccharide layer (EF116). A mutant (EF121) that was morphologically normal in appearance and had normal nitrogenase activity under microaerobic conditions (M. Murry and C. P. Wolk, unpublished observations) could have been defective in electron transfer to oxygen. Mutant EF104, also morphologically normal, appeared to be defective in the ability to transfer electrons to substrates of nitrogenase. In EF113, portions of the polysaccharide layer were seen only near the ends of the heterocysts: it is unclear whether only the polysaccharidesynthesizing apparatus was defective, or whether the entire program of differentiation aborted at an early stage, preventing a variety of differentiation-specific processes. Mutants EF114 and EF122, which arose in the same experiment and so may have been siblings, lacked the single abundant heterocyst-specific glycolipid of Anabaena sp. strain PCC 7120. The appearance, within the heterocyst envelope of these mutants, of a gap at the position normally occupied by the glycolipids suggests that molecules of glycolipid were not structurally excluded from the region that they would normally have occupied. The significance of mutant EF116 lies in the following facts. (i) Formation of the so-called fibrous layer of the heterocyst envelope is the first morphological indication of heterocyst differentiation at an ultrastructural level (15, 24). (ii) There is no clear evidence that this layer differs in composition from the polysaccharide that constitutes the great bulk of the outer layer of the heterocyst envelope. That is, synthesis of that polysaccharide mgy be one of the earliest, as well as one of the most massive, heterocystspecific biosyntheses. Because the envel pe polysaccharide of heterocysts in EF116 was less cohesive than that in the wild-type strain, we consider it probable that the polysaccharide itself was in some way structurally altered. The corresponding wild-type gene may ther0fore be under the control of a strong, development-speciflc promoter that is activated very early in differentiation. A-s such, it could be used to identify yet earlier genes that mediate the control, by available nitrogen, of heterocyst differentiation. If the akinetes of Anabaena sp. strain PCC 7120, like those of Anabaena cylindrica and Anabaena vqriabilis (1-3), have the same envelope polysaccharide as do the heterocysts of the respective species, such a mutation qould be used to test

whether the same genes are involved in the two distinctive differentiations. Two of the mutants isolated (strains EF113 and EF116) have been complemented with clones from a gene library of chromosomal DNA, derived from wild-type Anabaena sp. strain PCC 7120 and established in E. coli, in cosmid shuttle vector pRL25C. The clones were transferred to the cyanobacteria by means of RP-4-promoted conjugation (28). Fragments containing wild-type DNA corresponding to the genes mutated in those strains were further subcloned. Hitherto, the cloning of genes involved in N2 fixation in cyanobacteria relied upon hybridization with heterologous probes (19). Complementation of mutants of heterocystous cyanobacteria allows the cloning of cyanobacteria-specific, N2 fixation genes (e.g., genes related to the development of heterocysts), like those isolated in this work, for which heterologous probes cannot be available from other biological sources. The use of this same approach to carry out genetic analysis of other processes, such as oxygenic photosynthesis and chromatic adaptation, should be possible, because the RP4-dependent conjugative system (28) has been shown to work in several cyanobacteria, including facultatively heterotrophic strains and strains showing chromatic adaptation (7). Cosmid shuttle vector pRL25C can be used for such

FIG. 4. Complementation of mutant EF116 by a cosmid library from wild-type Anabaena sp. strain PCC 7120. The lawn of EF116 (photographed after 1 month of selection) has largely bleached, leaving dark green (against a background of pale orange spots) only those spots that have received, by conjugal transfer, a cosmid that complements the original mutation.

VOL. 170, 1988

MUTANTS OF ANABAENA SP. STRAIN PCC 7120

BE C R I I

II

AP

Av BgBgAp Nr

II I

I

\

I

43E7:

iI II

I

-10

-20

CC

i

I

BHBE I II

R Nd BE,No z \,,

I

I

'N I r

+10

C I

1243

C I

-1

+30

kb

41 Eli:

l:41B

FIG. 5. Restriction map of three cosmids that complement EF116. (Complementing cosmids 41C1 and 43G9 are very closely similar to 41B9 and 43E7, respectively.) A 2.8-kb subclone of cosmid 41B9 that also complements EF116 is approximately centered around the unique NruI site (here shown at 0 kb) in the three cosmids. Abbreviations: Ap, ApaI; Av, AvaII; BE, BstEII; Bg, BgII; BH, BssHII; C, ClaI; Nd, NdeI; No, NotI; Nr, NruI; R, EcoRI.

FIG. 6. Restriction

map

of cosmid 43G5 that complements

mu-

tant EF113. A 4.8-kb complementing subclone of this cosmid

includes the sole EcoRI site (at ca. 26 kb) of the insert. BE, BH, Nd, and R are as described in the legend to Fig. 5. Abbreviations: B, BamHI; Ba, BalI; Bl, BgIl; Eo, EcoO109; Na, NaeI; Nc, NcoI; M, MluI; P, PvuII; X, XmaIII. For the sake of simplicity, restriction sites present in the vector portion (thick band [Fig. 1]) of this cosmid and in the insert are here not shown in the vector portion, except for the EcoRI site, which is included to show the orientation of the vector.

studies, because it contains numerous potentially useful cloning sites for restriction endonucleases and (as a consequence of containing pDU1 from Nostoc sp. strain PCC 7524) can replicate in at least seven other filamentous cyanobacteria (27). Under our conditions of selection, the complementing cosmids can be recovered from the exconjugant cyanobacteria. Therefore, future complementation of mutants that have equally low frequencies of reversion may substitute a mixed pool of cosmids for the segregated library that we used. ACKNOWLEDGMENTS This work was supported by U.S. Department of Energy contract DE-AC02-76ERO-1338, by National Science Foundation grant PCM-8402500, and by U.S. Department of Agriculture grant 85CRCR-1-1560. E.F. was the recipient of a Fulbright-Ministerio de Jtducacion y Ciencia (Spain) Fellowship. G.S. was supported in part by a grant from the Max Kade Foundation, Inc. LITERATURE CITED 1. Cardemil, L., and C. P. Wolk. 1976. The polysaccharides from heterocyst and spore envelopes of a blue-green alga. Methylation analysis and structure of the backbones. J. Biol. Chem. 251:2967-2975.

2. Cardemil, L., and C. P. Wolk. 1979. The polysaccharides from heterocyst and spore envelopes of a blue-green alga. Structure of the basic repeating unit. J. Biol. Chem. 254:736-741. 3. Cardemil, L., and C. P. Wolk. 1981. Polysaccharides from the envelopes of heterocysts and spores of the blue-green algae Anabaena variabilis and Cylindrospermum licheniforme. J. Phycol. 17:234-240. 4. Currier, T. C., J. F. Haury, and C. P. Wolk. 1977. Isolation and preliminary characterization of auxotrophs of a filamentous cyanobacterium. J. Bacteriol. 129:1556-1562. 5. Fling, M. E., J. Kopf, and C. Richards. 1985. Nucleotide sequence of the transposon Tn7 gene encoding an aminoglycoside-modifying enzyme, 3"(9)-O-nucleotidyltransferase. Nucleic Acids Res. 13:7095-7106. 6. Flores, E., and G. Schmetterer. 1986. Interaction of fructose with the glucose permease of the cyanobacterium Synechocystis sp. strain PCC 6803. J. Bacteriol. 166:693-696. 7. Flores, E., and C. P. Wolk. 1985. Identification of facultatively heterotrophic, N2-fixing cyanobacteria able to receive plasmid vectors from Escherichia coli by conjugation. J. Bacteriol. 162: 1339-1341. 8. Grillo, J. F., P. J. Bottomley, C. Van Baalen, and F. R. Tabita. 1979. A mutant of Anabaena sp. CA with oxygen-sensitive nitrogenase activity. Biochem. Biophys. Res. Commun. 89: 685-693. 9. Harm, W. 1980. Biological effects of ultraviolet radiation. Cambridge University Press, Cambridge. 10. Haury, J. F., and C. P. Wolk. 1978. Classes of Anabaena variabilis mutants with oxygen-sensitive nitrogenase activity. J. Bacteriol. 136:688-692. 11. Herrero, A., J. Elhai, B. Hohn, and C. P. Wolk. 1984. Infrequent cleavage of cloned Anabaena variabilis DNA by restriction endonucleases from A. variabilis. J. Bacteriol. 160:781-784. (Errata, 162:858, 1985). 12. Herrero, A., and C. P. Wolk. 1986. Genetic mapping of the chromosome of the cyanobacterium, Anabaena variabilis. Proximity of the structural genes for nitrogenase and ribulosebisphosphate carboxylase. J. Biol. Chem. 261:7748-7754. 13. Hu, N.-T., T. Thiel, T. H. Giddings, and C. P. Wolk. 1981. Anabaena and Nostoc cyanophages from sewage settling ponds. Virology 114:236-246. 14. Lambert, J. A. M., E. Williams, P. A. O'Brien, and J. A. Houghton. 1980. Mutation induction in the cyanobacterium Gloeocapsa alpicola. J. Gen. Microbiol. 121:213-219. 15. Lang, N. J., and P. Fay. 1971. The heterocysts of blue-green algae. II. Details of ultrastructure. Proc. R. Soc. Lond. B. 178: 193-203. 16. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. 17. Nichols, B. W., and B. J. B. Wood. 1968. New glycolipid specific to nitrogen-fixing blue-green algae. Nature (London) 217: 767-768. 18. Peterson, R. B., and C. P. Wolk. 1978. High recovery of nitrogenase activity and of "Fe-labeled nitrogenase in heterocysts isolated from Anabaena variabilis. Proc. Nati. Acad. Sci. USA 75:6271-6275. 19. Rice, D., B. J. Mazur, and R. Haselkorn. 1982. Isolation and physical mapping of nitrogen fixation genes from the cyanobac-

1244

WOLK ET AL.

terium Anabaena 7120. J. Biol. Chem. 257:13157-13163. 20. Rippka, R., J. Deruelles, J. B. Waterbury, M. Herdman, and R. Y. Stanier. 1979. Generic assignments, strain histories and properties of pure cultures of cyanobacteria. J. Gen. Microbiol. 111:1-61. 21. Saito, N., and H. Werbin. 1970. Purification of a blue-green algal deoxyribonucleic acid photoreactivating enzyme. An enzyme requiring light as a physical cofactor to perform its catalytic function. Biochemistry 9:2610-2620. 22. Simon, R. D. 1978. Survey of extrachromosomal DNA found in the filamentous cyanobacteria. J. Bacteriol. 136:414-418. 23. Spence, D. W., and W. D. P. Stewart. 1987. Heterocystless mutants of Anabaena PCC7120 with nitrogenase activity. FEMS Microbiol. Lett. 40:119-122. 24. Wilcox, M., G. J. Mitchison, and R. J. Smith. 1973. Pattern formation in the blue-green alga Anabaena. II. Controlled

J. BACTERIOL. proheterocyst regression. J. Cell Sci. 13:637-649. 25. Wilcox, M., G. J. Mitchison, and R. J. Smith. 1975. Mutants of Anabaena cylindrica altered in heterocyst spacing. Arch. Microbiol. 103:219-223. 26. Wolk, C. P. 1982. Heterocysts, p. 359-386. In N. G. Carr and B. A. Whitton (ed.), The biology of cyanobacteria. Blackwell Scientific Publications, Ltd., Oxford. 27. Wolk, C. P., E. Flores, G. Schmetterer, A. Herrero, and J. Elhai. 1985. Development of the genetics of heterocyst-forming cyanobacteria, p. 491-496. In H. J. Evans, P. J. Bottomley, and W. E. Newton (ed.), Nitrogen fixation research progress. Martinus Nijhoff Publishers, Dordrecht. 28. Wolk, C. P., A. Vonshak, P. Kehoe, and J. Elhai. 1984. Construction of shuttle vectors capable of conjugative transfer from Escherichia coli to nitrogen-fixing filamentous cyanobacteria. Proc. Natl. Acad. Sci. USA 81:1561-1565.