d-Amino Acids Trigger Biofilm Disassembly

See allHide authors and affiliations

Science  30 Apr 2010:
Vol. 328, Issue 5978, pp. 627-629
DOI: 10.1126/science.1188628


Bacteria form communities known as biofilms, which disassemble over time. In our studies outlined here, we found that, before biofilm disassembly, Bacillus subtilis produced a factor that prevented biofilm formation and could break down existing biofilms. The factor was shown to be a mixture of d-leucine, d-methionine, d-tyrosine, and d-tryptophan that could act at nanomolar concentrations. d-Amino acid treatment caused the release of amyloid fibers that linked cells in the biofilm together. Mutants able to form biofilms in the presence of d-Amino acids contained alterations in a protein (YqxM) required for the formation and anchoring of the fibers to the cell. d-Amino acids also prevented biofilm formation by Staphylococcus aureus and Pseudomonas aeruginosa. d-amino acids are produced by many bacteria and, thus, may be a widespread signal for biofilm disassembly.

Most bacteria form multicellular communities known as biofilms in which cells are protected from environmental insults (1, 2). However, as biofilms age, nutrients become limiting, waste products accumulate, and it is advantageous for the biofilm-associated bacteria to return to a planktonic existence (2). Thus, biofilms have a finite lifetime, characterized by eventual disassembly. Bacillus subtilis forms communities on semi-solid surfaces and thick pellicles at the air/liquid interface of standing cultures (1, 35). Cells in the biofilms are held together by an extracellular matrix consisting of exopolysaccharide and amyloid fibers composed of the protein TasA (57). The exopolysaccharide is produced by the epsA-O operon, and the TasA protein is encoded by the yqxM-sipW-tasA operon (8). After 3 days of incubation in a biofilm-inducing medium, B. subtilis formed thick pellicles at the air/liquid interface of standing cultures (Fig. 1A). Upon incubation for an additional 3 to 5 days, however, the pellicles lost their integrity (Fig. 1B). To investigate whether mature biofilms produce a factor that triggers biofilm disassembly, we asked whether a conditioned medium would prevent pellicle formation when added to a fresh medium (9). Medium from an 8-day-old culture was applied to a C18 column (Sep Pak, Waters, Milford, MA), and concentrated eluate from the column was added to a freshly inoculated culture. The eluate was sufficient to prevent pellicle formation (Fig. 1C). Concentrated eluate from a 3-day-old culture had little effect on pellicle formation (Fig. 1D). Further purification of the factor was achieved by eluting the cartridge stepwise with methanol. Elution with 40% methanol resulted in a fraction that was active in inhibiting pellicle formation (Fig. 1E), but had little effect on cell growth (fig. S1). The activity was resistant to heating at 100°C for 2 hours and proteinase K treatment (Fig. 1F).

Fig. 1

Conditioned medium blocks pellicle formation. B. subtilis strain NCIB3610 was grown at 22°C in 12-well plates in a liquid biofilm-inducing medium for 3 days (A) or 8 days (B). (C and D) Cells grown for 3 days in a medium with added dried and resuspended methanol eluate (1:100 v/v) from a C18 column (Sep Pak) that had been loaded with conditioned (cond.) medium from a 6- to 8-day-old culture (C) or a 3-day-old culture (D). The final concentration of concentrated factor added to the wells represented a 1:4 dilution on a volume basis of the original conditioned medium. (E) The factor was further purified on the C18 column by stepwise elution with methanol. The result of adding 3 μl of the 40% methanol eluate is shown. (F) Before addition to fresh medium, the 40% methanol eluate was incubated with proteinase K beads. (G to L) Wells containing MSgg medium supplemented with d-tyrosine (3 μM), d-leucine (8.5 mM), l-tyrosine (7 mM), l-leucine (8.5 mM), d-tyrosine (3 μM), and d-alanine (10 mM) or with d-leucine (8.5 mM) and d-alanine (10 mM) were inoculated with strain NCIB3610 and incubated for 3 days. The diameter of all wells was 22.1 mm.

Bacteria produce d-amino acids in stationary phase (10). We asked whether the biofilm-inhibiting factor was composed of one or more d-amino acids. Indeed, d-tyrosine, d-leucine, d-tryptophan, and d-methionine were active in inhibiting biofilm formation in a liquid medium, as well as on a solid medium (Fig. 1, G and H, and figs. S2 and S3). In contrast, the corresponding l-isomers and d-isomers of other amino acids (such as d-alanine and d-phenylalanine) were inert in our biofilm-inhibition assay. Next, we determined the minimum concentration needed to prevent biofilm formation. Individual d-amino acids varied in their activity, with d-tyrosine being more effective (3 μM) than d-methionine (2 mM), d-tryptophan (5 mM), or d-leucine (8.5 mM). A mixture of all four d-amino acids was particularly potent, with a minimum inhibitory concentration of ~10 nM. Thus, d-amino acids can act synergistically. Not only did the d-amino acids prevent biofilm formation, but they also disrupted existing biofilms. The addition of d-tyrosine or a mixture of the four d-amino acids (but not the corresponding l-amino acids) caused pellicle breakdown (Fig. 2A).

Fig. 2

d-amino acids break down pellicles, and inhibition of pellicle formation by conditioned medium depends on racemase genes. (A) Three-day-old cultures with added (as a 10 μl drop to the surface of the pellicles to achieve the indicated final concentration) l-tyrosine (7 mM), a mixture of l-amino acids (5 mM each), d-tyrosine (3 μM), or a mixture of d-amino acids (2.5 nM each). (B) Effect of concentrated C18 column (Sep Pak) eluate from a conditioned medium from an 8-day-old culture from the wild type or from a strain (IKG55) doubly mutant for ylmE and racX. Scale bars, 7 mm.

d-Amino acids are generated by racemases (11). Genetic evidence consistent with the idea that the biofilm-inhibiting factor is d-amino acids came from mutants of ylmE and racX, genes whose predicted products exhibit sequence similarity to known racemases. Strains mutant for ylmE or racX alone showed a modest delay in pellicle disassembly (fig. S4). However, pellicles formed by cells doubly mutant for the putative racemases were substantially delayed in disassembly (fig. S4). Conversely, cells engineered to overexpress ylmE were blocked in biofilm formation (fig. S5). Also, conditioned medium from the double mutant was ineffective at inhibiting biofilm formation, in contrast to conditioned medium from the wild type (Fig. 2B). Next, we asked whether d-amino acids are produced in sufficient abundance to account for disassembly of mature biofilms. Accordingly, we carried out liquid chromatography–mass spectrometry, followed by the identification of the d-amino acids using derivatization with Nα-(2,4-dinitro-5-fluorophenyl)-l-alaninamide (l-FDAA) on conditioned medium collected at early and late times after pellicle formation. d-tyrosine (6 μM), d-leucine (23 μM), and d-methionine (5 μM) were present at concentrations at or above those needed to inhibit biofilm formation by day 6, but only at concentrations of <10 nM at day 3. In contrast, the ylmE racX double mutant was blocked in d-tyrosine production and impaired in d-leucine production at day 6 (table S1).

How do d-amino acids disassemble biofilms? d-amino acids did not inhibit growth (fig. S6), nor did they inhibit the expression of the matrix operons epsA-O and yqxM-sipW-tasA (fig. S7). d-amino acids are incorporated into the peptide side chains of peptidoglycan in place of the terminal d-alanine (10). Using 14C– d-tyrosine, we confirmed that d-tyrosine (but not 14C–l-proline) was incorporated into the cell wall (fig. S8), with incorporation beginning at day 3 (fig. S9). Finally, in keeping with the idea that d-amino acids act via their incorporation into the wall, the effects of d-tyrosine and other d-amino acids were prevented by d-alanine (Fig. 1, K and L, and fig. S2).

We hypothesized that TasA fibers are anchored to the cell wall and that the incorporation of biofilm-disassembling d-amino acids into the cell wall might disengage the fibers from their anchor. To investigate this possibility, we examined the localization of a functional fusion of TasA with the fluorescent reporter mCherry. Treatment with d-tyrosine had little effect on the accumulation of TasA-mCherry (fig. S10). In contrast, when the cells were washed by centrifugation, re-suspended, and then examined by fluorescence microscopy, untreated cells, which were often in clumps, were intensely decorated with TasA-mCherry (Fig. 3A). In contrast, d-tyrosine–treated cells, which were largely unclumped, showed only low levels of fluorescence (17-fold lower; table S2). Similar results were obtained with d-leucine and with the d-amino acid mixture. We also carried out electron microscopy (EM) with gold-labeled antibodies to TasA (anti-TasA) to visualize unmodified TasA. TasA fibers were anchored to the cells of untreated pellicles (Fig. 3B, images 1 and 2). In contrast, cells treated with d-tyrosine consisted of a mixture of cells that were largely undecorated with TasA fibers and free TasA fibers or aggregates of fibers that were not anchored to cells (Fig. 3B, images 3 to 6).

Fig. 3

d-tyrosine causes the release of TasA fibers. (A) Cell association of TasA-mCherry by fluorescence microscopy. Wild-type cells and yqxM6 (IKG51) mutant cells containing the tasA-mCherry fusion were grown to stationary phase (optical density at 600 nm = 1.5) with shaking in a biofilm-inducing medium in the presence (+d-Tyr) or absence (untreated) of d-tyrosine (6 μM), as indicated. The cells were then washed in phosphate-buffered saline and visualized by fluorescence microscopy. Scale bars, 4 μm. (B) Cell association of TasA fibers by EM. 24-hour-old cultures were incubated without (images 1 and 2) or with (images 3 to 6) d-tyrosine (0.1 mM) for an additional 12 hours. TasA fibers were stained by immunogold labeling with anti-TasA antibodies and visualized by transmission electron microscopy. Solid black arrows indicate fiber bundles; open arrows indicate individual fibers. Scale bars in the enlargements of images 2 and 4 and in image 6, 100 nm; all other scale bars, 500 nm. Images 1 and 2 show fiber bundles attached to cells; images 3, 4, and 6 show individual fibers and bundles detached from cells; and images 3 to 5 show cells with little or no fiber material. For further details, see the supporting online material (9).

Next, we isolated d-amino acid–resistant mutants (Fig. 4A). Wrinkled papillae appeared spontaneously on the flat colonies formed during growth on a solid medium containing d-tyrosine (Fig. 4A) or d-leucine (fig. S2). When purified, these spontaneous mutants gave rise to wrinkled colonies and pellicles in the presence of individual d-amino acids. We isolated several such mutants and found that they contained mutations in the 3′ region of yqxM (table S3). Two mutations that conferred resistance to d-tyrosine were examined in detail: (i) yqxM2 was an insertion of G:C at base pair 728 and (ii) yqxM6 was a deletion of A:T at base pair 569 (Fig. 4B). The presence of yqxM2 and yqxM6 restored clumping and cell decoration by TasA-mCherry to cells treated with d-tyrosine (Fig. 3A and fig. S12; see above text). Because YqxM is required for the association of TasA with cells (6), the discovery that the biofilm-inhibiting effect of d-amino acids could be overcome by mutants of YqxM reinforces the view that the effect of d-amino acid incorporation into the cell wall is to impair the anchoring of the TasA fibers to the cell. A domain near the C terminus of YqxM could trigger the release of TasA in response to the presence of d-amino acids in the cell wall.

Fig. 4

Biofilm formation by YqxM mutants resistant to d-tyrosine. (A) Wild-type or yqxM6 (IKG51) mutant cells grown on a solid (images 1 to 3) or liquid (images 4 to 6) biofilm-inducing medium. Cells were grown in the absence (1 and 4) or presence (2, 3, 5, and 6) of d-tyrosine (3 μM). Arrows in image 2 indicate papillae of spontaneous resistances mutants. Scale bars, 3 mm. (B) C-terminal amino acid sequences for wild-type YqxM and the indicated frame-shift mutants (altered sequences in blue).

Finally, we wondered whether d-amino acids would inhibit biofilm formation by other bacteria. The pathogens Staphylococcus aureus and Pseudomonas aeruginosa form biofilms on plastic surfaces (12), which can be detected by washing away unbound cells and staining the bound cells with crystal violet. d-tyrosine and the d-amino acid mixture were effective in preventing biofilm formation (fig. S13), whereas l-tyrosine and l-amino acids had no effect. Furthermore, the effect of d-amino acids was prevented by the presence of d-alanine (fig. S13), suggesting that d-amino acids acted to block biofilm formation by replacement of d-alanine in the peptide side chain. Given that many bacteria produce d-amino acids, these amino acids may provide a general strategy for biofilm disassembly. If so, then d-amino acids might prove widely useful in medical and industrial applications for the prevention or eradication of biofilms.

Supporting Online Material

Materials and Methods

Figs. S1 to S13

Tables S1 to S3


References and Notes

  1. Materials and methods are available as supporting material on Science Online.
  2. We thank T. Norman for the analysis of table S2. I.K.-G. and D.R. are postdoctoral fellows of the Human Frontier Science Program and Fulbright/Ministerio de Educación y Ciencia (Spain), respectively. This work was funded by NIH grants to R.K. (GM58213), J.C. (GM086258 and CA24487), and R.L. (GM18546), as well as grants from BASF to R.K. and R.L. Additionally, we have filed a patent application on the use of d-amino acids for biofilm disassembly.
View Abstract

Stay Connected to Science

Navigate This Article