Contribution of Cell Elongation to the Biofilm Formation of Pseudomonas aeruginosa during Anaerobic Respiration

Pseudomonas aeruginosa, a gram-negative bacterium of clinical importance, forms more robust biofilm during anaerobic respiration, a mode of growth presumed to occur in abnormally thickened mucus layer lining the cystic fibrosis (CF) patient airway. However, molecular basis behind this anaerobiosis-triggered robust biofilm formation is not clearly defined yet. Here, we identified a morphological change naturally accompanied by anaerobic respiration in P. aeruginosa and investigated its effect on the biofilm formation in vitro. A standard laboratory strain, PAO1 was highly elongated during anaerobic respiration compared with bacteria grown aerobically. Microscopic analysis demonstrated that cell elongation likely occurred as a consequence of defective cell division. Cell elongation was dependent on the presence of nitrite reductase (NIR) that reduces nitrite (NO2 −) to nitric oxide (NO) and was repressed in PAO1 in the presence of carboxy-PTIO, a NO antagonist, demonstrating that cell elongation involves a process to respond to NO, a spontaneous byproduct of the anaerobic respiration. Importantly, the non-elongated NIR-deficient mutant failed to form biofilm, while a mutant of nitrate reductase (NAR) and wild type PAO1, both of which were highly elongated, formed robust biofilm. Taken together, our data reveal a role of previously undescribed cell biological event in P. aeruginosa biofilm formation and suggest NIR as a key player involved in such process.


Introduction
Pseudomonas aeruginosa propagates as complex, highly organized communities known as biofilms in the environment and during various infections [1]. Bacterial biofilms are the most common causes of chronic P. aeruginosa infection and are difficult to eradicate. For example, P. aeruginosa in biofilms have been reported to be more resistant to H 2 O 2 [2], a range of antibiotics [3] and various heavy metals [4]. Moreover, bacteria grown as biofilms are more resistant to neutrophil-mediated host defenses than are their free-living planktonic counterparts [5].
P. aeruginosa can generate sufficient energy even under anaerobic conditions through respiration using nitrate (NO 3 2 ) or nitrite (NO 2 2 ) as terminal electron acceptors [6,7]. It was demonstrated that (i) the oxygen potential of abnormally altered CF airways, which are highly susceptible to chronic P. aeruginosa infection, is extremely low [8] and (ii) that nitrate (NO 3 2 ) and nitrite (NO 2 2 ) are present in large quantities inside patient airways [9,10]. Given the fact that PAO1, when grown under anaerobic conditions, becomes more resistant to antibiotic treatment [11], these findings provided a novel insight that reflects P. aeruginosa infection dynamics in CF airways and suggested to change the way of confronting P. aeruginosa airway infection.
Genome-wide microarray analysis revealed that the expression of a total of 691 genes (12% of the genome) was modulated upon anaerobic growth demonstrating that P. aeruginosa, as a versatile organism, can actively adapt itself to growth with alternative electron acceptors [12]. Cellular events that specifically occur during anaerobic respiration include decreased production of pyocyanin [13] and elastase [14], increased levels of alginate secretion [8], consistent production of sublethal levels of NO [6,15,16] and enhanced biofilm formation [17,18]. Although the molecular basis of these anaerobiosis-induced changes is not fully understood, it is likely that P. aeruginosa may increase its survival fitness inside the patient airway, especially at the chronic stage, by reducing the production of virulence factors and increasing the capability to form biofilm. Consistent with this notion, P. aeruginosa isolates from chronically infected patients exhibited weaker elastase activity than those recovered from colonized CF patients or from pediatric patients without CF [19] and possessed mutations in lasR gene [20].
Biofilm formation is a developmental process by which bacteria undergoes significant phenotypic and genetic changes including the acquisition of antibiotic resistance and modulation of growth properties [21,22,23]. It is of particular interest that P. aeruginosa forms more robust biofilm under anaerobic respiration [17], because this demonstrates a resistant mode of bacterial proliferation under a condition that specifically represents the abnormal CF airway. Among many determinants that contribute to biofilm maturation, initial surface attachment [17,24] and secretion of matrix molecules [25] are often considered to be critical. In this study, a unique morphological change that occurs specifically under anaerobic growth and thus influences biofilm formation was observed for the first time in P. aeruginosa. Wild type PAO1 was highly elongated during anaerobic respiration and an elongationdefective mutant formed very weak biofilm. Cell elongation was attributable to the organism's response to nitric oxide (NO). Ultimately, this observation may form a basis for deciphering the mechanisms by which P. aeruginosa modulates its pathogenic properties under anaerobic growth condition.

Bacterial strains and growth conditions
Wild type P. aeruginosa strain PAO1 and mucoid strain FRD1 have been previously described [26]. P. aeruginosa transposon insertion mutants, DnarG, DnirS, DnorC and Danr mutant strains were purchased from a P. aeruginosa transposon mutant library (http://www.genome.washington.edu/UWGC/pseudomonas) and sequence verified. L-broth (LB, 10 g tryptone, 10 g NaCl, 5 g yeast extract per liter) was used to grow bacteria. Anaerobic bacterial growth was achieved using a GasPak anaerobic system (Becton, Dickinson and Company, Franklin Lakes, NJ) or in a Coy anaerobic chamber (Coylab Inc. Grass Lake, MI). The gas composition inside the anaerobic chamber was a trimix of nitrogen, hydrogen, and carbon dioxide (90, 5, and 5%, respectively). Dry anaerobic indicator strip (Becton, Dickinson and Company) was used to confirm the generation of anaerobic atmosphere inside the jar. In addition, no detectable growth of PAO1 in plain LB was confirmed in each anaerobic experiment. To support anaerobic growth, KNO 3 or NaNO 2 (Sigma-Aldrich) was added to the medium. When required, pH of the LB medium was adjusted to pH 7.8 using 50 mM phosphate.

Scanning Electron Microscope and cell length measurement
Bacterial cells were visualized by scanning electron microscope (SEM) and confocal microscope. For the sample preparation of SEM, bacterial suspension was fixed with PBS containing 2% glutaraldehyde and 0.1% paraformaldehyde for 2 hrs and stained with 1% OsO 4 . Samples were then coated with gold by an ion sputter (IB-3 Eiko, Japan) and examined with a scanning electron microscope (FE SEM S-800, Hitachi, Japan) at an acceleration voltage of 20 kV. Images were processed with ESCAN 4000 software (Bummi Universe Co., LTD, Seoul, Korea). For the cell length measurement, more than 100 straight-lined cells were randomly chosen in the digitized SEM images and distance between two ends was automatically calculated.

Confocal microscopy
Differential Interference Contrast (DIC) images were acquired using a confocal laser scanning microscope (FV-1000; Olympus Optical Co. Ltd., Japan) and its operating software, FV10-ASW (ver. 02.01). Prior to the image analysis, aliquots of bacterial cultures were washed with PBS and mounted in wells of 8-well Lab-Tek TM chambered coverglass (cat. no. 155411, Nalge Nunc International, Rochester, NY). After scanning with 488 nm laser at a sampling speed of 12.5 ms/pixel, a 6406640 pixel, 12-bit image (57.51 mm6 57.51 mm) was acquired. UPLSAPO 100XO (Olympus) objective lens was used for the bacterial cell image analysis. Images were saved as a TIF file with embedded 5 mm scale bar.
For the membrane visualization, PAO1 grown in LB plus 0.4% NO 3 2 to the early stationary phase under aerobic or anaerobic condition was washed with PBS and the cells were resuspended in 200 ml of PBS containing 10 mM of the lipophilic membrane dye TMA-DPH (Invitrogen Corp. Carlsbad, CA). Again, 8-well Lab-Tek TM chambered coverglass (cat. no. 155411, Nalge Nunc International) was used to mount samples on the objective lens. Image acquisition was conducted as described for the DIC image analysis except that samples were scanned at 405 nm and emissions were collected at 461 nm.
For the nucleoid staining, Syto 9 green fluorescent dye (Invitrogen Corp. Carlsbad, CA) was used at 10 mM final concentration. To capture the green fluorescence, samples were scanned at 488 nm and emission was detected through a 520 nm band filter. The DIC and green fluorescence images were collected simultaneously.
Counting the cluster formation in the planktonic culture of P. aeruginosa strains To count the number of clusters, 20 ml of each bacterial culture was pipetted onto the slide glass. A 22640 mm coverslip (Paul Marienfeld GmbH & Co. KG, Lauda-Koenigshofen, Germany) was then placed over the rectangular area of the slide glass. Nikon SE optical microscope (Nikon Vision Co., Ltd) was used to view and count the number of clusters in the sample. Cluster counting was performed while the objective lens was manually moved from the top left to the bottom right region of the coverslip. Aggregates that consist of more than 10 cells were considered as clusters.

Biofilm assays
The ability of P. aeruginosa to develop biofilm was assessed with a modified microtiter plate assay as described previously [24]. Briefly, P. aeruginosa aerobic preculture grown in LB was inoculated (1:100 dilution) into the anaerobic culture media (i.e. LB containing NO 3 2 or NO 2 2 ) in 24-well or 96-well plates and incubated anaerobically for 18 hrs at 37uC without agitation. Biofilm formed by P. aeruginosa strains was stained with 0.1% crystal violet (CV) and the stained CV was dissolved in 95% ethanol for measurement of the absorbance at 540 nm. Because the capacity to form biofilm is proportional to the bacterial cell growth, OD 540 nm was normalized with cell mass determined by measuring OD 600 nm . To test the effect of antibiotic treatment on the biofilm formation, sub-MIC concentrations of carbenicillin, tobramycin or ciprofloxacin were added to the aerobic PAO1 culture in 96-well plates. The minimal inhibitory concentration (MIC) of selected antibiotics was determined as described previously [27].

Quantitative real time-PCR (qRT-PCR) analysis
Aliquots (1.5 ml) of bacterial cell cultures were harvested by spinning down at 14,000 rpm for 5 min. Pelleted cells were resuspended with 1 ml of Trizol (Invitrigen) and then 0.2 ml of chloroform was added. After incubating 5 min at room temperature, Trizol-chloroform mixture was centrifuged for 10 min at 4uC to separate the aqueous phase containing RNA. The rest of RNA purification steps were carried out using RNeasy kit (Qiagen) following the manufacturer's instruction. The extracted RNA samples were subjected to PCR to verify the absence of contaminating DNA. The resulting RNA samples were quantified using a Nanodrop spectrophotometer (model no. ASP2680, CellTAGen Inc., Seoul Korea). For cDNA synthesis, 2 mg RNA template was mixed with 1 ml of 100 pmoles/ ml random primer (59-NSNSNSNSNS-39, where N = A,T,C, or G and S = C or G) and dNTP mix in 15 ml total volume. The mixture was then treated at 65uC for 5 min, followed by 5 min incubation on ice. Next, one unit of Primescript reverse transcriptase (Takara Bio Inc., Shiga, Japan) was added with 1 ml of 5x primescript reaction buffer and 13 ml DEPC-treated water. The mixture was then incubated at 42uC for 1 hr and at 70uC for 15 min. Real time PCR reaction was monitored using StepOne Real-time PCR system (Applied Biosystems, Carlsbad, CA). For the reaction, SYBR premix Ex Taq (Takara) was used following the manufacturer's instructions. Primer sets used to amplify cDNA are listed in Table S1. The PCR cycle was 95uC for 10 min and 40 cycles of 95uC for 20 s and 60uC for 20 s, followed by 95uC for 15 s and 60uC for 1 min. Transcript levels of rpoD gene were similar in cells grown by either aerobic or anaerobic respiration and thus, used for the normalization.
Statistical analysis. Data are expressed as mean 6 SEM (standard error of mean). An unpaired Student's t-test was used to analyze the data. A p-value of ,0.05 was considered statistically significant. All the experiments were repeated for reproducibility.

Results
Cell elongation occurs in PAO1 grown by anaerobic respiration using NO 3

as an alternative electron acceptor
Understanding cell biological features of P. aeruginosa that occur specifically upon anaerobic respiration would provide a better insight into the bacterial pathogenic mechanisms under such condition. To address this important question, we first investigated cellular morphology of PAO1 grown aerobically or anaerobically by scanning electron microscope. When grown in LB with aeration, PAO1 exhibited normal rod shape morphology with a cell length of ,1.2 mm (Fig. 1A). The addition of 0.4% NO 3 2 to the culture medium did not cause any detectable change in cell shape (Fig. 1B). In contrast, PAO1 grown by NO 3 2 respiration under anaerobic condition was highly elongated (Fig. 1D) compared to aerobically grown cells. Software-aided cell length measurement clearly demonstrated that cells are highly elongated upon anaerobic NO 3 2 respiration (,5.2 mm vs. ,1.2 mm, Fig. 1E). Cell elongation did not occur in PAO1 incubated anaerobically for the same period of time in growth medium that lacked NO 3 2 (Fig. 1C). No discernable growth was observed in this particular culture further proving that the presence of alternative electron acceptor is crucial for anaerobic growth in P. aeruginosa (Fig. S1, hatched bar). In addition, no similar elongation was observed in anaerobic cultures of mutant bacteria disrupted in nitric oxide (NO) reductase (DnorCB, Fig. 1G) or ANR, one of the master anaerobic transcriptional regulators (Danr, Fig. 1H). These two mutant strains were defective in anaerobic respiration and thus no anaerobic growth was detected in previous works [6,28]. It was also reported that P. aeruginosa can support, to a lesser extent, anaerobic growth by arginine fermentation [29]. As shown in Fig. 1I, no elongation was observed in cells grown by arginine fermentation. Together, these results suggest that cell elongation is caused by an active bacterial response to anaerobic Cell elongation is likely caused by defective cell division Fig. 1D revealed that most of elongated cells are free of invaginated cell wall pattern, which may indicate that the formation of septal peptidoglycan [30] is suppressed in PAO1 grown by anaerobic respiration. To gain a better idea of the membrane structure of elongated cells, we visualized bacterial cells stained with TMA-DPH, a lipophilic dye that specifically binds to cell membrane and thus outlines the cell periphery [31]. As shown in Fig. 2B, PAO1 grown by anaerobic respiration was highly elongated compared with that grown by aerobic respiration ( Fig. 2A) further validating the results described in Fig. 1. It is of interest that continuous staining throughout the entire cell envelope was observed in elongated bacteria as indicated with red arrows (Fig. 2B). Moreover, no clear pattern of cleavage furrow formation was detected in the same cell images. This result strongly suggests that normal cellular machinery for the cell division is affected in these anaerobically grown bacteria.
It was reported that inhibition of chromosomal DNA replication by UV irradiation caused a cell elongation phenotype in E. coli [32]. Therefore, we sought to examine whether or not the replication of chromosomal DNA occurred normally in such elongated cells. To address this issue, we stained the cells with syto 9, a cell-permeant green fluorescence dye that specifically stains nucleic acid. In aerobically grown and thus rod-shaped cells, fluorescent signal was detected in the entire area of cell further suggesting that bacterial genome is not limited within a membrane-enclosed suborganelle (Fig. 3A, O 2 ). Interestingly, multiple segregating nucleoids were observed in most of elongated cells (Fig. 3B, 2O 2 ). When we compared identical fields of images acquired by DIC or syto 9 staining side by side, distinctly segregated nucleoids were detected in elongated cells (red arrows in Fig. 3B). Together, these image analyses strongly suggest that cell elongation is due not to defects in DNA replication or chromosome segregation, but to the incomplete septum formation.
Next, we compared the transcript levels of representative genes involved in cell division and cell wall synthesis by quantitative real time PCR (qRT-PCR). We first analyzed zipA (PA1528), ftsZ (PA4407) and ftsA (PA4408), because formation of Z-ring, which requires the cooperative assembly of FtsZ, FtsA and ZipA at the mid-cell region is essential for the initiation of cell division process [33]. Transcript level of rpoD gene, which exhibited almost identical level of expression under aerobic or anaerobic growth, was used as a normalization control. As shown in Fig. 4, expression of zipA gene was ,3-fold lower in PAO1 grown by anaerobic respiration than in aerobically grown PAO1. A more than 10-fold decrease was observed in transcript levels of both ftsZ and ftsA during anaerobic vs. aerobic respiration. We then examined transcript levels of murD (PA4414) and murF (PA4416) genes encoding UDP-N-acetylmuramoylalanine-D-glutamate ligase and UDP-N-acetylmuramoylalanyl-D-glutamyl-2,6-diaminopimelate-D-alanyl-D-alanyl ligase, two important enzymes involved in peptidoglycan synthesis. Likewise, transcript levels of these two genes in PAO1 grown by anaerobic respiration were only ,13% of those achieved in cells grown by aerobic respiration. This result suggests that expression of genes involved in Z-ring formation and peptidoglycan synthesis, two critical steps for the optimal cell division, is significantly reduced during the anaerobic growth in P. aeruginosa.

Functional nitrite reductase (NIR) is required for anaerobiosis-triggered cell elongation
Anaerobic respiration in P. aeruginosa involves a sequential reduction of nitrate (NO 3 2 ) or nitrite (NO 2 2 ) to N 2 [6]. Higher cell yield was obtained in PAO1 growth using NO 3 2 as an electron acceptor than NO 2 2 -supported anaerobic growth (Fig.  S1), suggesting that NO 3 2 reduction is the major step that is coupled to ATP synthesis in wild type P. aeruginosa. Since cell elongation only occurred in anaerobically respiring P. aeruginosa, the anaerobic respiration pathway was further dissected to uncover the crucial step that is responsible for cell elongation. To address this issue, cell elongation phenotypes of DnarG and DnirS mutant, which became defective in NAR or NIR, respectively, were tested (Fig. 5A). Consistent with findings described in Fig. 1, wild type strain PAO1 and two mutant strains grown by aerobic respiration maintained their regular rod-shape morphology even in the presence of 15 mM NO 3 2 (Fig. 5B, E and H). PAO1 anaerobically grown on NO 3 2 (Fig. 5C) or NO 2 2 (Fig. 5D) was invariably elongated compared to its aerobically grown counterpart (Fig. 5B). This suggests that anaerobic growth supported by either electron acceptor can trigger cell elongation in P. aeruginosa. To minimize the growth-inhibitory effect of NO 2 2 on bacterial growth [26], the pH of the culture medium was adjusted to 7.8, when NO 2 2 is used as an electron acceptor. It was important to note that the DnarG and DnirS mutants did not grow on NO 3 2 or NO 2 2 , respectively, due to the lack of the enzyme that can reduce the corresponding electron acceptor (Fig. S1) and thus, no cell elongation was observed in each of these two anaerobic cultures ( Fig. 5F and J, * denotes no growth). This result further validates that active respiratory growth is the prerequisite for the cell elongation during anaerobic growth. Importantly, DnarG mutant bacteria grown by NO 2 2 respiration were as elongated as its parental strain, PAO1 (Fig. 5G), whereas DnirS mutant cells were not elongated upon anaerobic growth on NO 3 2 (Fig. 5I). Final cell density in these two cultures was comparable to each other (Fig.  S1), suggesting that the formation of these two contrasting cell shapes (Fig. 5G vs. 3I) was not caused by the differential bacterial cell growth.
Moreover, mucoid P. aeruginosa, FRD1, determined to possess undetectable NIR activity [26], was not elongated under the same NO 3 2 respiring condition (Fig. S2). Together, these results suggest that NIR-mediated reduction of NO 2 2 to NO is the critical step that triggers anaerobiosis-induced cell elongation in P. aeruginosa.
Biofilm formation of the non-elongated DnirS mutant was significantly reduced under the NO 3 2 respiring condition It has been previously shown that during anaerobic respiration, P. aeruginosa formed significantly more robust biofilm compared to when bacteria grow aerobically [17,18]. This finding was successfully reproduced in our crystal violet biofilm staining assay as shown in Fig. 6A. Since elongated cell morphology was only observed under anaerobic respiration condition, we sought to examine if this enhanced biofilm formation is caused by a modified cell biological feature associated with cell elongation. To address this question, the biofilm formation of DnarG and DnirS mutants that showed distinct cell elongation phenotypes were compared ( Fig. 5G and I). pH-buffered L-Broth media was used to maximize bacterial growth by NO 2 2 respiration. Wild type PAO1 formed very robust biofilm in both NO 3 2 and NO 2 2 -stimulated anaerobic growth (Fig. 6B). Quantification analysis of biofilm formation by measuring OD 540 nm /OD 600 nm , however, indicated that denser biofilm was formed under the condition of NO 3 2 respiration than NO 2 2 respiration (Fig. 6C). This was likely due to greater bacterial growth by NO 3 2 respiration (Fig. S1). It is important to note that the DnarG mutant, which was elongated during anaerobic growth using NO 2 2 as an electron acceptor (Fig. 5G), formed biofilm that was almost as robust as the PAO1 biofilm under the same growth conditions (Fig. 6B). In contrast, biofilm formation was completely abrogated in the non-elongated DnirS mutant upon NO 3 2 respiration (Fig. 6B). Again, this distinct biofilm formation was not due to the differential growth of these two anaerobic cultures, because cell growth of the DnirS mutant on NO 3 2 was slightly greater than the DnarG mutant in NO 2 2 (Fig.  S1). As expected, anaerobic cell growth of DnarG on NO 3 2 or DnirS on NO 2 2 was minimal and thus, no biofilm was formed in each of these two cultures. These results suggest that anaerobiosisinduced cell elongation, which is dependent on the presence of NIR activity, plays a critical role in robust biofilm formation under these conditions.

Cell elongation and biofilm formation were both suppressed in the presence of Carboxy-PTIO, a stoichiometric scavenger in PAO1
Although nitric oxide (NO), the product of NIR, is further reduced to N 2 O, steady-state level of nitric oxide (NO) is maintained during anaerobiosis [16,34]. Since (i) NIR activity, which generates NO as its product, is required for cell elongation as shown in Fig. 5, and (ii) the enzymatic action of NIR is the only source of NO production under anaerobic growth conditions [6], it was postulated that production of intracellular NO may account for the anaerobiosis-induced cell elongation. To address this notion, the cell elongation phenotype of PAO1 grown in the presence of 2 mM carboxy-PTIO, a stoichiometric NO scavenger was tested. In our previous work, carboxy-PTIO, which penetrates the periplasm and scavenges NO before it escapes to the cytoplasm or is reduced by NOR, successfully protected P. aeruginosa strains from NO-mediated intoxication [17]. As shown in Fig. 7A and B, cell elongation was suppressed in the presence of carboxy-PTIO (4.8761.89 mm vs. 2.3560.54 mm). Fig. 6 indicate that cell elongation is an important cellular event that contributes to the biofilm formation under the condition of anaerobic respiration. To further prove the effect of cell elongation on biofilm formation, biofilm formation in a condition where cell elongation was hindered by the addition of carboxy-PTIO was examined. As shown in Fig. 7C, a dosedependent decrease in biofilm formation was detected in the presence of increasing amount of carboxy-PTIO. Upon growth with 2 mM carboxy-PTIO, biofilm robustness was ,47% of that of the control anaerobic biofilm suggesting that the perturbation of cell elongation exerted an adverse effect on biofilm formation. No adverse effect of added carboxy-PTIO on cell viability was observed up to 2 mM concentration (data not shown).

Clump formation was induced during anaerobic growth
Importantly, we often observed highly cohesive cell clusters in the elongated P. aeruginosa. Fig. 8A shows a representative image of clusters formed by PAO1 during an anaerobic planktonic culture. When we attempted to quantify the cluster formation in anaerobic vs. aerobic cultures of PAO1, ,30 (614) clusters per 20 ml of overnight culture that was mounted on the slide glass were identified in our microscopic analysis. Similar clusters were detected neither in the aerobic P. aeruginosa cultures nor anaerobic DnirS mutant culture (data not shown). This data strongly suggest that cell elongation that specifically occurs during anaerobic respiration is involved in the cohesive clump formation that eventually leads to the robust biofilm formation.
Two distinct genetic loci, pel and psl produce carbohydrate-rich biofilm matrix that can hold bacterial cells together [35,36,37]. Therefore, it is of interest to examine whether Pel and/or Psl polysaccharide matrix could contribute to the anaerobiosisinduced clump formation. As a way to address this issue, we performed qRT-PCR on pelE and pslB, genes representing each of these two clusters. Shown in Fig. 8B is the transcriptional modulation of these two genes under aerobic vs. anaerobic growth conditions. Transcription of pelE gene was not changed upon anaerobic growth, while the transcript level of pslB gene was decreased to ,20% of what was observed in cells grown aerobically. This result suggested that levels of Pel and/or Psl polysaccharides produced during anaerobic respiration would not be higher than those produced during aerobic growth and thus, anaerobiosis-induced enhancement of clump formation is not likely to be associated with altered production of extracellular polysaccharides.

Biofilm formation was promoted in the presence of carbenicillin that induced cell filamentation
Our results suggest that robust anaerobic biofilm formation is likely mediated by modified cellular features associated with anaerobiosis-specific cell elongation in P. aeruginosa. To provide further evidence that supports the correlation between cell elongation and biofilm formation in P. aeruginosa, we tested if biofilm formation can also be promoted by other stimulus that causes cell elongation. Because cell elongation occurs upon treatment with carbenicillin in P. aeruginosa [38,39,40], we compared biofilm formation of PAO1 treated with carbenicillin versus ciprofloxacin and tobramycin. As shown in Fig. 9, biofilm formation of PAO1 was significantly increased in the presence of sublethal doses of carbenicillin. In our microscopic analysis, a high degree of cell filamentation was also observed confirming results from aforementioned referenced studies (data not shown). It is of note that PAO1 was highly resistant to carbenicillin and MIC was determined to be 750 mg/ml, a value consistent with previous findings [40,41]. In contrast, only a mild increase in biofilm formation was observed in PAO1 treated with tobramycin or ciprofloxacin (Fig. 9). No cell shape change was observed in aerobic growth with each of these two antibiotics (data not shown). This result demonstrated that treatment with an antibiotic that specifically induced cell elongation also promoted biofilm formation further suggesting the positive influence of cell elongation on biofilm formation in P. aeruginosa.

Discussion
Biofilm is a microbial community grown as an aggregate or on a surface with distinct architecture [23]. Biofilm research using P. aeruginosa as a model organism has been performed using in vitro biofilm grown under aerobic respiration. Although the presence of local regions with reduced oxygen potential has been proposed to exist inside bacterial biofilm [42,43], previous works showed that anaerobically growing P. aeruginosa formed significantly robust biofilm compared with bacteria growing by aerobic respiration [17,18]. This observation is of clinical importance because the airway of chronic CF patients was suggested to be anaerobic due to the accumulation of abnormally thickened and viscous mucus on top of the airway epithelium [8]. In addition, anaerobically growing P. aeruginosa exhibited higher resistance to a range of currently used antibiotics than their aerobically grown counterparts [11], further suggesting that P. aeruginosa proliferates inside the patient airway by employing two different modes of antibioticresistant growth, i.e. biofilm and anaerobiosis.
The molecular mechanisms behind the enhanced biofilm formation under anaerobic growth conditions are not clearly defined. Here, a unique morphological feature of P. aeruginosa grown by anaerobic respiration was identified for the first time and its effect on biofilm formation was investigated. Our results revealed that anaerobic growth of P. aeruginosa is concurrently accompanied by abnormally altered cell division. PAO1 grown during anaerobic respiration was ,5-fold more elongated than aerobically grown cells by a mechanism associated with defective cell division. In P. aeruginosa, cell elongation was reported to be caused by nutrient deprivation [44]. It was postulated that bacteria elongate to increase their nutrient uptake as a part of their adaptation process for starvation. However, it seems unlikely that anaerobiosis-induced cell elongation occurred for a similar reason since (i) the cells were grown in rich media (i.e. L-Broth) and (ii) the density at which the culture was harvested for the microscopic image was as low as OD 600 of ,0.5. In addition, NO 3 2 -supported anaerobic growth is considerably luxuriant and thus the final density of an anaerobic culture of wild type PAO1 supplemented with 100 mM NO 3 2 was almost comparable with that of aerobic culture (data not shown).
Cell elongation (or filamentation) was also reported to take place under nonpermissive conditions, such as high growth temperature [45], the treatment with certain antibiotics [46] and the UV irradiation [32]. But, bacterial growth under these highly stressful conditions was completely ceased and furthermore, the level of cell elongation appeared to be significantly greater than what was observed in the present study. This suggests that although P. aeruginosa favors the utilization of oxygen to generate energy and can grow faster aerobically, cells are not placed under any stress condition in the absence of oxygen and thus, the anaerobiosisinduced cell elongation occurs within a range that is permissive for cell growth.
Our subsequent experiments uncovered the fact that cell elongation was mediated by NO, produced consistently as a spontaneous intermediate of anaerobic respiration [16,17]. Our previous work [6] and a work by Cork and Poole using E. coli as a test organism [47] indicated that NIR activity was necessary to produce NO under anaerobic conditions. Consistent with this  finding, the DnirS mutant that was devoid of NIR activity was not elongated under NO 3 2 respiration conditions (Fig. 5). These findings were further proven by the observation that NO-triggered cell elongation was suppressed by the addition of membranepenetrable carboxy-PTIO (Fig. 7).
During airway infection, P. aeruginosa encounters significant level of neutrophils infiltrated into the airway mucus. Along with macrophages and airway epithelial cells, neutrophils produce NO as an important defense molecule [48]. Because our results suggest that elongated cells form more robust biofilm, an interesting hypothesis would be whether P. aeruginosa responds to host-derived NO during the infectious process in order to change its cell shape. To test this idea, we have attempted to see the effect of exogenously added NO via acidified NO 2 2 [26] on the cell shape change in PAO1 under aerobic growth condition, but no morphological change in response to NO was observed (data not shown). This may suggest that NO-triggered cell elongation only occurs under anaerobic respiration condition using endogenously produced metabolic NO. More experiments are necessary to precisely determine the effect of NO on the bacterial cell shape change.
Our results also demonstrated that the elongated cells are inclined to form highly cohesive clumps, which we believe  accounts for the robust biofilm formation under anaerobic condition (Fig. 8). Because clump formation was not observed in non-elongated mutant strains (DnirS mutant) and aerobically grown P. aeruginosa, this finding suggests that anaerobiosis-induced cell elongation likely caused changes in membrane properties that result in the clump formation. Potential questions to be addressed in the future include; (i) how different the cell wall composition of elongated cells is from rod-shaped P. aeruginosa and (ii) how the endogenously produced NO can signal a mechanism by which cells undergo elongation.
The discovery of anaerobiosis-induced morphological change inspired the present study to test its effect on biofilm formation. Since (i) cell-to-cell contact, which is an important determinant for biofilm formation, can be facilitated by an enlarged cell surface and (ii) highly cohesive autoaggregates were only detected in elongated cells, it was hypothesized that the robust anaerobic biofilm formation might result from the elongated cell shape. To address this important question, the elongation-defective mutant(s) were subjected to isolation. However, a genome-wide mutant library screen to isolate such mutants was not feasible, due to the lack of an efficient assay system that would allow us to isolate the defective mutants. Alternatively, the effect of the disruption of genes involved in the anaerobic respiration pathway on the elongation phenotype was examined since cell elongation occurred as a consequence of anaerobic respiration in wild type PAO1 (Fig. 1). The results described in Fig. 5 demonstrated that the DnirS mutant remained rod-shaped, while the DnarG mutant was elongated as PAO1 under anaerobic growth. This finding provided us with an opportunity to compare the biofilm forming capability of the mutants that were in marked contrast to each other in their cell morphology.
NIR was also reported to be required for the type III secretion system and virulence towards human monocyte cell line THP-1 and Caenorhabditis elegans even in conditions where oxygen is present [7,49]. This suggests that NIR is likely important in P. aeruginosa pathogenesis irrespective of the environmental oxygen concentration. Results presented in the current study, however, clearly demonstrated the role of NIR in anaerobiosis-specific condition. Our conclusion that enhanced biofilm formation occurred under anaerobic condition was due, at least in part, to the NIR-mediated cell elongation and was based on the following results; (i) the non-elongated DnirS mutant did not form biofilm under NO 3 2 respiration conditions (Fig. 6), (ii) the DnarG mutant grown by NO 2 2 respiration was highly elongated and formed very robust biofilm (Fig. 6), and (iii) biofilm formation was decreased by the addition of carboxy-PTIO that inhibited cell elongation (Fig. 7).
In summary, this report provides initial insights into the cell shape-dependent biofilm-forming characteristics of P. aeruginosa. Anaerobic biofilm formation that represents a persistent bacterial survival mechanism inside the patient airway requires the activity of NIR. Therefore, it is suggested that this key player could serve as a target, inhibition of which can be beneficial as novel antipseudomonas therapeutic interventions are further developed. Figure S1 Final cell density of anaerobic cultures of three P. aeruginosa strains. Bacterial strains were grown for 18 hours in LB +15 mM NO 3 2 (black bars) or NO 2 2 (gray bars) anaerobically. For a negative control, PAO1 was grown in plan LB. * p,0.01 vs. growth with NO 3 2 . (TIF)