JOURNAL OF BACTERIOLOGY, Dec. 1987, p. 5877-5879 Vol. 169, No. 120021-9193/87/125877-03$02.00/0Copyright X 1987, American Society for Microbiology

Mechanism of Attachment of Swarm Cells of Thiothrix niveaJOHN M. LARKIN* AND RODNEY NELSON

Department of Microbiology, Louisiana State University, Baton Rouge, Louisiana 70803

Received 23 June 1987/Accepted 30 August 1987

Swarm cells of Thiothrir nivea were found to possess a group of fimbriae at one pole. The other pole eitherwas bare or possessed from one to three fimbriae. By using this polarity as a marker, it was found that theinitial step in attachment of swarm cells involves the fimbriated pole and that this initial step is followed by theproduction of holdfast material.

Thiothrix nivea lives in a gradient in the environment,existing in flowing water in which the sulfide concentration isabout 0.1 to 1.0 mg/liter, the oxygen concentration is about10% or less of saturation, and the pH is near neutrality (1, 2,5, 6, 9). In a suitable environment, T. nivea cells grow aslong filaments attached to a substrate or as rosettes orfilaments attached by their poles to objects or other orga-nisms (including themselves). Single cells, called swarmcells, are released from the open end of the sheath and

suitable environment and escapes being swept away by thecurrent (9). The mechanism of this attachment is unknownand is the subject of this report.

T. nivea JP2, which was isolated (7) and described previ-ously (8), was used in this study. Stock cultures weremaintained on either MP or MY broth (8) over a plug of thecorresponding agar in screw-cap tubes. For experimentalcultures, the organism was grown in MY broth in whichthiosulfate was substituted for sulfide.

Ilm

FIG. 1. Negative stain of a T. nivea JP2 swarm cell. The arrows point to fimbriae that are located on a single pole.

exhibit gliding motility when on a solid surface (8). Swarmcells are responsible for dispersal of the colony; they mayattach to a substrate and produce a multicellular filament, orthey may attach to other swarm cells and produce a rosette.The attachment of T. nivea to a surface or to other cells, asin a rosette, is the mechanism by which it remains in a

* Corresponding author.

Negative stains on Formvar-coated grids were preparedwith phosphotungstic acid, pH 7.4. Cells for scanning elec-tron microscopy were fixed in 1 M Veronal acetatebuffer-3% glutaraldehyde, dehydrated in a graded ethanolseries, dried to the critical point, and coated in a sputtercoater as described previously (12).

Negatively stained preparations were scanned in the elec-tron microscope until a well-isolated swarm cell was. found,and each pole of the cell was examined. Only single cells

5877

5878 NOTES

FIG. 2. Negative stain of the fimbriae on a pole of a swarm cellat high magnification. tohFIG. 3. Negative stain of two swarm cells which have attachedathighmagnification. ~~~~~~~toeach other at the fimbriated pole.

were examined because they would be newly producedswarm cells from the pole of a Thiothrix filament (9). If bothends of the cell were not visible, the cell was not examined.A total of 51 individual swarm cells were found in which

both ends of each cell could be examined. Of these cells, 43(84%) had many fimbriae at one end, while the other end was

bare (Fig. 1). Four swarm cells (8%) had many fimbriae atone pole and only one to three fimbriae at the other pole.Four cells (8%) had no detectable fimbriae. Thus, 47 of 51swarm cells (92%) exhibited an easily observed polarity,which could be detected by the location of their fimbriae.Neither pole of a swarm cell or unattached filament had anaccumulation of holdfast material, which would have madeobservation difficult. At high magnification, the fimbriaecQnsisted of single thin filaments or filaments which were

aggregated into much larger filaments (Fig. 2). The polarityof the swarm cell was used as a marker to determine whichend of a swarm cell harbored the attachment mechanism.Swarm cells that were attached at their poles in groups of

two to five cells were located on negatively stained prepara-tions (Fig. 3). The presence and location of fimbriae on eachcell of these developing rosettes were determined. Singlecells that were attached to the sides of filaments were

examined in the same way.Frequently, it was not possible to see the fimbriae on cells

that were attached because of the accumulation of holdfastmaterial. However, because of the polarity of the swarm

cells it was possible to determine whether attachment oc-

curred at the fimbriated or nonfimbriated end by examiningthe free end of the cell, which had no holdfast material.The distribution of fimbriae on the unattached ends of cells

in developing rosettes is shown in Table 1. Of the poles awayfrom the attachment sites, 91% had no fimbriae, 6% had one

to three fimbriae, and 3% had a group of fimbriae. Thus, 97%of the swarm cells in a beginning rosette attached at thefimbriated pole. Of swarm cells that attached to the sides ofa filament, 100% attached with the heavily fimbriated pole.Overall, 98% of all of the attached swarm cells attached withthe end that had the tuft of fimbriae.These data demonstrate that the fimbriated pole of the T.

nivea swarm cell is always at the point of attachment.Additional evidence that fimbriae are involved in attachmentwas found when, after several years in culture, this strain ofT. nivea simultaneously lost both the ability to producefimbriae and the ability to attach to a substrate.

Holdfast material was not seen on unattached swarm cells,making observation of the poles a simple matter. Fimbriaecould sometimes be seen between the poles of the cells innewly attached pairs. As the number of swarm cells in a

rosette increased, it became increasingly difficult to see theattachment area because of the accumulation of slimelikematerial. Thus, it appears that holdfast material is producedafter the initial attachment by the fimbriae. In a scanningelectron micrograph of the center of a rosette (Fig. 4),strands of slimelike material can be seen extending betweencells, and they undoubtedly play some role in adherence.The mechanism(s) used by bacteria to produce rosettes

may differ among species, but several surface structureshave been implicated. Moore and Marshall (11) examinedrosette formation in a Hyphomicrobium sp. and suggestedthat the polar flagellum may be involved in attachment. An

TABLE 1. Distribution of fimbriae on the free ends of swarm cells that had attached to other cells

No. of free ends' in developing rosette having: Total no. of free No. of single cells Total no. ofType of free endsa 2 Attached 3 Attached 4 Attached 5 Attached ends examined attached to a ends examined

cells cells cells cells (% of total) filament (% of total no.)

Observed 18 3 6 5 32 (100) 14 46 (100)With 0 fimbriae 15 3 6 5 29 (91) 10 39 (85)With 1 to 3 fimbriae 2 0 0 0 2 (6) 4 6 (13)With tuft of fimbriae 1 0 0 0 1 (3) 0 1 (2)

a The end away from the attachment site.The ends of cells in some rosettes could not be seen and were not counted.

J. BACTERIOL.

NOTES 5879

FIG. 4. Scanning electron micrograph of the center of a T. nivea

rosette. The arrows point to strands of slimelike material.

acidic polysaccharide was responsible, at least in part,for attachment and rosette formation in Asticcacaulis bipros-thecum; fimbriae and flagella were not involved (13).MacRae et al. (10) examined 22 gliding bacteria and found

that 19 of them had a tuft of fimbriae at one pole, while theother pole was bare. In a follow-up study (3), they concludedthat fimbriae are organelles which function to establish andmaintain intercellular contacts The communal swarmingbehavior of myxobacteria requires the presence of fimbriae(3, 4).Our results indicate that the swarm cells of T. nivea are

similar to those of the gliding bacteria described by MacRaeet al. (10) in that they have a tuft of fimbriae at one pole onlyand the fimbriae appear to act as structures used to establishintercellular contact.

LITERATURE CITED1. Bahr, H., and W. Schwartz. 1956. Untersuchugen zur Okologie

farbloser fadiger Schwefelmikroben. Biol. Z. 75:451-464.2. Bland, J. A., and J. T. Staley. 1978. Observations on the biology

of Thiothrix. Arch. Microbiol. 117:79-87.3. Dobson, W. J., H. D. McCurdy, and T. H. MacRae. 1979. The

function of fimbriae in Myxococcus xanthus. II. The role offimbriae in cell-cell interactions. Can. J. Microbiol. 25:1359-1372.

4. Kaiser, D. 1979. Social gliding is associated with the presence ofpili in Myxococcus xanthus. Proc. Natl. Acad. Sci. USA 76:5952-5956.

5. Keil, F. 1912. Beitrage zur Physiologie der farblosen Schwefel-bakterien. Beitr. Biol. Pflanzen. 11:335-372.

6. Lackey, J. B., E. W. Lackey, and G. B. Morgan. 1965. Taxon-omy and ecology of sulfur bacteria, vol. 19, p. 3-23. Eng. Prog.Report no. 119. University of Florida, Gainesville.

7. Larkin, J. M. 1980. Isolation of Thiothrix in pure culture andobservation of a filamentous epiphyte on Thiothrix. Curr. Mi-crobiol. 4:155-158.

8. Larkin, J. M., and D. L. Shinabarger. 1983. Characterization ofThiothrix nivea. Int. J. Syst. Bacteriol. 33:841-846.

9. Larkin, J. M., and W. R. Strohl. 1983. Beggiatoa, Thiothrix,and Thioploca. Annu. Rev. Microbiol. 37:341-367.

10. MacRae, T. H., W. J. Dobson, and H. D. McCurdy. 1977.Fimbriation in gliding bacteria. Can. J. Microbiol. 23:1096-1108.

11. Moore, R. L., and K. C. Marshall. 1981. Attachment and rosetteformation by hyphomicrobia. Appl. Environ. Microbiol.42:751-757.

12. Strohl, W. R., and J. M. Larkin. 1978. Cell division andtrichome breakage in Beggiatoa. Curr. Microbiol. 1:151-155.

13. Umbreit, T. H., and J. L. Pate. 1978. Characterization of theholdfast region of wild-type cells and holdfast mutants ofAsticcacaulis biprosthecum. Arch. Microbiol. 118:157-168.

VOL. 169, 1987

As