APPLIED AND ENVIRONMENTAL MICROBIOLOGY, May 1992, p. 1629-1635
Vol. 58, No. 5
0099-2240/92/051629-07$02.00/0
Measurements of the Distribution of Adenylate Concentrations and Adenylate Energy Charge across Pseudomonas aeruginosa Biofilms S. L. KINNIMENT* AND J. W. T. WIMPENNY School of Pure and Applied Biology, University of Wales College Cardiff CFI 3TL, United Kingdom
Cardiff,
Received 16 December 1991/Accepted 27 February 1992
Adenine nucleotide pools and adenylate energy charge distributions were determined by using a laboratorygenerated quasi-steady-state Pseudomonas aeruginosa biofilm. The method used involved freezing and sectioning of the intact biofilm, followed by extraction and assay of the adenylates in the sectioned material. Results indicated an increase in adenylate energy charge of about 0.2 units from the bottom to the surface of the biofilm. However, energy charge values were generally low throughout the biofilm, reaching a maximum of only 0.6 units. Of the adenylates measured, AMP was the predominant nucleotide, especially in the deeper parts of the biofilm profile. Biofilm is the general term that applies to microbial communities forming coherent layers on solid surfaces. These structures are ubiquitous and can lead to serious economic loss to industry, besides posing health risks in the dental and medical spheres. These problems are exacerbated because biofilms are generally less sensitive to antimicrobial agents than are freely suspended organisms (22). As a spatially heterogeneous community, very little is known about the structure and physiology of biofilm. A number of experimental models which allow the investigation of biofilm throughout the growth cycle exist (10, 24, 28). The constant-depth film fermentor developed by Coombe and colleagues (5) and by Peters and Wimpenny (23) generates quasi-steady-state biofilm. This is defined on the basis that total protein and viable count remain constant over a period of time. Biofilm can be of any chosen thickness up to about 500 p,m. One indicator of the energetic status of living cells is the adenylate pool (1). Sensitive assays which are capable of measuring femtomole quantities of ATP exist (27). It was felt that this level of sensitivity would allow us to measure adenylates in 12-p,m-thick cryostat sections of biofilm. Adenine nucleotides, especially ATP, play a central role as intermediate carriers of chemical energy linking catabolism and biosynthesis. In vitro studies have revealed that the activities of certain enzymes are affected by the concentration of ATP; others are affected by ADP or AMP. The adenylate energy charge (ECA) is a measure of the effect of the ratios of adenosine phosphate concentrations on the rate of cellular metabolism (12) and is defined as follows: ECA = ([ATP] + 1/2[ADP])/([ATP] + [ADP] + [AMP]). The ECA iS a linear measure of the total amount of potential energy momentarily stored in the adenine nucleotide pool. It is a dimensionless number which ranges from 0 to 1 (2). Since there is claimed to be a positive correlation between the growth potential or functional capacity of a cell and its ECA (4), these measurements provide a framework for estimating metabolic potentials of naturally occurring microbial populations. This article describes a method for freezing and sectioning biofilm generated in the constant-depth film fermentor. By *
using this technique, adenylate pools and energy charge values have been determined as a function of depth through a Pseudomonas aeruginosa biofilm.
MATERIALS AND METHODS Organisms. A metalworking fluid isolate, identified as P. aeruginosa (14), was used to inoculate the fermentors. This microorganism was selected because it emerged as the predominant bacterium in the development of a stable biofilm derived from contaminated metalworking fluids. The microorganism was stored by freezing on glass beads at -70-C (11). Media. The amine-carboxylate medium used was a simplified simulated cutting fluid based on a mixture of fatty acids (predominantly oleic acid) and triethanolamine supplemented with mineral salts and vitamins. The medium contained, per liter of distilled water, an 0.085% (vol/vol) mixture of cutting fluid components consisting, by weight, of 60% triethanolamine and 40% fatty acids (of which 66% was oleic acid) and (in grams) the mineral salts FeSO4. 7H20 (1.5 x 10-2), MgSO4. 7H20 (0.2), ZnSO4- 7H20 (1.75 x 10-3),
MnSO4 * 4H20 (1 X
10-4), CUS04
5H20 (1 X
10-4),
NH4Cl (1.0), K2HPO4 3H20 (1.2), and CaCl2- 6H20 (1 x 10-2) and the vitamins inositol (1 x 10-2), calcium pantothenate (2.0 x 10-3), and biotin (1.0 x 10-5). For plate cultures, the same medium was used, except that the cutting fluid components were increased to 2% (vol/vol) and 2% (wt/vol) agar (Difco) was added. Cutting fluid components were supplied by Castrol Research Laboratories (Whitchurch Hill, Pangbourne, Reading, United Kingdom). Other chemicals were AnalaR grade (BDH) unless otherwise stated. Biofilm growth. Biofilm formation. The constant-depth film fermentor used by Kinniment and Wimpenny (14) was used to generate a reproducible biofilm. The constant-depth film fermentor consists of an enclosed rotating stainless steel disc in which are located 15 removable biofilm pans. Each biofilm pan is drilled with a central threaded hole, and around the edge of this are arranged six biofilm plugs. The latter are recessed to the selected depth up to 500 p.m by using a specially machined tool. Biofilm develops on the flat surface of the plug in the recessed space. The steel disc rotates beneath angled polytetrafluoroethylene (PTFE) scraper
Corresponding author. 1629
1630
KINNIMENT AND WIMPENNY
blades which remove excess growth, maintaining the biofilm at a constant depth. Medium drips onto the steel disc directly in front of one scraper blade. The latter distributes it over the biofilm pans, and any excess is wiped away. Samples of biofilm can be taken by replacing pans at selected time intervals by using two sterilizable stainless steel tools. A threaded extractor tool is used to remove and replace the pans, while a flat tamping tool is used to push replacement pans down flush within the turntable. The fermentor is heat sterilizable, the design allowing aseptic removal and replacement of biofilm pans. In addition, the system is operated under controlled gas phase and temperature conditions. The inoculum for the fermentor was prepared by aseptically flooding a 48-h-streaked amine-carboxylate plate of P. aeruginosa with 2 ml of sterile Tris-HCl buffer, pH 8.8. One milliliter of the homogenized mixture was then removed and resuspended in 4 ml of sterile buffer. This inoculum was allowed to recirculate through the fermentor in 500 ml of medium for 12 h. The main 10-liter batch of sterile medium was then aseptically connected to the fermentor. At the same time, the waste outlet port was connected to a large sterile waste receiver. The fermentor was operated with a medium flow rate of 1 ml min-1, an air flow rate of 250 ml min-1, a disc rotation speed of 3 rpm, and a constant temperature of 30°C. Sample pans were taken from the fermentor at selected time intervals. Each pan contained six PTFE sample plugs recessed to a depth of 300 p.m. The diameter of the plug surface was 4.75 mm, giving a surface area of approximately 17.72 mm2 and a biofilm volume, if the available space was completely full, of about 5.32 mm3. The fermentor was run to quasi-steady state before samples were removed for sectioning. The fermentor was closed down after 10 liters of medium had passed through the system. Estimation of total protein accumulation. Protein analysis based on a modified Lowry method was carried out (18). A sample pan was removed and four plugs were carefully pushed out of the pan, with the biofilm still intact on the surface. Protein on each plug was assayed separately by being boiled in 0.5 ml of 1 M NaOH for 5 min. These were each diluted fourfold in 1 M NaOH, and 0.5 ml (each) of the diluted mixtures was then used in the standard Folin procedure. Protein standards containing 5 to 150 p.g of protein (bovine albumin fraction five; BDH) were prepared each time the assay was performed. The absorbance values for the biofilm samples were then read against these standard curves. Viable count. The two remaining sample plugs from each pan were removed with flame-sterilized forceps. To disperse the biofilm, the plugs were transferred to 10 ml of sterile deflocculant solution in a universal container. The deflocculant consisted of 0.01% (wt/vol) Cirrasol (ICI Organics) in 0.01% (wt/vol) sodium pyrophosphate as described by Gayford and Richards (8). Approximately 250 2.5- to 3.5-mmdiameter glass beads (BDH) had been added to this solution. Sample plugs and biofilm were then agitated by vortex mixing for 90 s. After treatment, the sample was serially diluted in Tris-HCl buffer at pH 8.8 and 0.1 ml of each dilution was plated out in triplicate onto amine-carboxylate plates. The plates were incubated at 30°C for 24 to 48 h, and CFU were determined. Method for sectioning sampled biofilm. Before sectioning, a method for freezing and mounting the biofilm to a cryostat sample holder was developed, and it is described below. Agar (2% [wt/vol]) (Oxoid) plates were poured aseptically
APPL. ENVIRON. MICROBIOL.
to a depth of about 5 mm. Agar plugs were cut from these plates with a sterile no. 15-size cork borer. The plug of agar was removed from the petri dish with a sterile spatula, inverted, and placed on the upper surface of the agar remaining in the petri dish. A circle of lens tissue was placed on top of the agar plug, and a drop of water was placed on it to moisten it. With forceps, a cryostat sample holder was touched firmly to the surface of the damp lens tissue. The whole assembly was inverted and the sample holder, with the agar layer attached, was then lifted away from the surface of the agar plate and placed in a polystyrene container. Sample pans containing quasi-steady-state biofilm were removed from the fermentor at 55 and 105 h. The PTFE sample plugs were carefully removed from the sample pans. Sample plugs, biofilm upper surface down, were positioned between a pair of forceps which were then placed in the jaws of a micromanipulator. The whole assembly was placed directly above the sample holder, and 1 drop of water was quickly placed onto the surface of the agar. By using the micromanipulator, the PTFE sample plug with biofilm attached was lowered onto the drop of water so that the biofilm surface was just submerged. Liquid nitrogen was then poured into the polystyrene container to the depth of the agar plug, and the agar and biofilm were left to freeze. Once the biofilm was frozen, the forceps were removed from the micromanipulator and the PTFE sample plug was pulled away from the frozen biofilm. A modification of this method was used for one set of replicate samples; the main difference was that the sample holder was chilled to -37°C prior to being mounted onto the agar plug. The biofilm, previously frozen in liquid nitrogen, was then stuck onto the frozen agar surface with a drop of water, and once frozen the PTFE sample plug was removed. The time between taking samples of and freezing the biofilm was approximately 2 to 3 min for the first method and 1 to 2 min for the second. The sample holder was fixed to the microtome in a Starlet 2212 cryostat (Bright Instrument Company Ltd., Huntingdon, United Kingdom), and the knife blade was adjusted to cut 12-p.m, entire, horizontal biofilm sections. Sections were removed with a clean sterile microspatula. In the first method, sections were removed in threes; in the second method, the first section was discarded and subsequent sections were then removed in pairs. A total of six biofilm profiles were sectioned, three 55-h and three 105-h samples. Method of adenylate extraction of sectioned biofilm. The method described below was adapted from that used by Lundin and Thore (19) and Scourfield (25). Each set of two or three sections was transferred to 0.5 ml of 2.3 M perchloric acid containing 6.7 mM EDTA. The mixture was stirred constantly for 15 min on ice and centrifuged for 3 min in an Eppendorf bench top centrifuge at approximately 13,000 x g, and the supernatant was neutralized with 2 M KOH-0.5 M triethanolamine buffer. The mixture was left on ice for 5 min until all the potassium perchlorate had precipitated and then was centrifuged for a further 2 min. The supernatant was finally frozen at -70°C until needed. Adenylate and energy charge analyses of sectioned biofilm. ATP, ADP, and AMP were assayed in triplicate by standard methods (19, 20) by using a purified firefly lantern extract (Bio-Orbit Oy, Turku, Finland). An LKB 1251 computerinterfaced luminometer (Bio-Orbit Oy) was used for the bioluminescence assay. One hundred twenty-five microliters of the assay mixture was placed in a cuvette in the sample holder of the luminometer. Once the sample had been loaded
ADENYLATE MEASUREMENTS THROUGH BIOFILM
VOL. 58, 1992
(a)
suggests a doubling time of approximately 3.5 h
1-1 go
._: 0
I
10 0
25
_
I
50
75 100 Time (h)
125
75
125
150
17''5
lo 9
8.
X.
10
8
1 7
.0I
(P.max = 0.2
h-1) (Fig. lb). Analysis of adenylate concentration through the biofilm.
1000
(b)
1631
0I
52 10 6. 0
I
I
25
50
I
.
100
I
150
175
Time (h) FIG. 1. (a) Typical growth curve (each value expressed as miof protein per plug) of the biofilm. Error bars show standard deviations based on four sample plugs. (b) Viable count, expressed as CFU per plug (average for two sample plugs). crograms
into the detection chamber, 100 ,ul of ATP monitoring reagent (Bio-Orbit Oy) was rapidly injected, the sample and reagent were pulse-mixed, and the 10-s integral of light
emitted from each assay was recorded immediately. All assays were performed at 25°C. ATP was measured directly; ADP was determined after conversion to ATP by using pyruvate kinase (Sigma) and phospho(enol)pyruvate (Sigma); AMP was measured by being converted to ADP and then to ATP by using adenylate kinase (Sigma) together with the pyruvate kinase and phospho(enol)pyruvate. Internal standards were assayed to detect possible inhibition of the bioluminescence reaction, and an ATP calibration curve was also constructed.
The results of adenylate measurements showed random fluctuations (Fig. 2a and b); therefore, the data were smoothed by taking a three-point moving average. Although the first and last datum points are lost by using this approach, trends in adenylate levels and ECA values are easier to discern. Errors for the moving average adenylate concentrations and the ECA values were estimated by calculating the root mean square deviation of the original data from the moving average data for the former and then the ratios of random variables for the latter (13). An estimate of the probability of any significant difference in ECA between the base and surface points was calculated by comparing two samples, and then standard Student's t tables were consulted. Adenylate concentrations for two of the six biofilm profiles are shown in Fig. 2, as examples. For the remaining replicate samples, the total adenylates and the percentages of ATP, ADP, and AMP in relation to the total adenylates are presented in Table 1. The results for total adenylates and individual adenylates and the fraction of each expressed as a percentage of the total adenylates through the biofilm profile show similar trends for all six profiles (Fig. 2c to f and Table 1). Different biofilms were used for each profile, so some variation is to be expected. In the 55-h biofilms, concentrations of total adenylates, AMP, and ATP rose from the base and reached a maximum near the middle of the biofilm. By 105 h these peaks had moved toward the surface. ADP stays relatively constant, showing a small peak at the center of the 55-h biofilms and near the surface of the 105-h biofilms. All adenylate levels fall very near the surface of the biofilm in both sets of samples (Fig. 2c and d and Table 1). In the 55-h biofilms, the percentage of AMP in relation to the total adenylates drops toward the surface of the biofilm, that of ATP increases, and that of ADP stays relatively constant. For the 105-h biofilms, ATP and AMP follow a similar course but ADP either increases or decreases through the center before levelling off again near the surface. The dominant adenylate appeared to be AMP in all samples (Fig. 2e and f and Table 1). ATP levels were generally low. Values ranged from about 1 to 30 pmol per section. While the 55-h biofilms appeared to have reached a quasi-steady state, judging by the protein and CFU values, the number of sections taken from the 55-h samples was generally smaller than that of the 105-h samples. (Fig. 1 to 3). Examination of ECA through the biofilm. Energy charge values showed a consistent increase across both sets of biofilm from base to surface. While the range was not large, it represented a value of about 0.2, from values of 0.22 to 0.28 near the base to 0.4 to 0.45 near the surface of the 55-h biofilms (Fig. 3a) and 0.17 to 0.35 near the base to 0.37 to 0.6 at the surface of the 105-h biofilms (Fig. 3b). The probability of a significant difference between the points at the biofilm base and the surface ranged from 70 to 99.9%, with an average value of 90%.
RESULTS
Biofilm growth. The P. aeruginosa biofilm was grown in amine-carboxylate medium for a period of 150 h. Quasisteady-state conditions were reached after about 45 h. Quasi-steady-state protein values plateaued at about 250 jig per plug (Fig. la). The viable count was about 2 x 108 CFU per plug. The slope of the log of viable count versus time
DISCUSSION A number of different sectioning procedures were tested before the methods described in this article were selected. It was not always easy to produce complete sections, and occasionally material was lost during the transfer of sections from the microtome blade to the sample container. The raw
KINNIMENT AND WIMPENNY
1632
APPL. ENVIRON. MICROBIOL.
(c)
(a)
(e)
55 h moving average data
55 h orginal data
1
au
C)
* .2 5)
5).r
70
base
60
4+
50 Cu 40 I 30 0 20 10
2,
55 h moving average data -
surface
+ =
-
0I 0
(b)
10 20 Section number
30
0
10
20 Section number
30
0
(D
105 h original data
10 20 Section number
30
105 h moving average data
120
70- base
rA100. n080 0 ° 80
0-
4
44
c -%
50-
1
cAf4 10 'e
surface
40V 30o 2010 -
53
0 Section number
Section number
I
n
10 20 Section number
30
FIG. 2. Graphical examples of adenylate concentrations through the biofilm. (a and b) Concentration of adenylates in each section (original data) (error bars show standard deviations of triplicate samples). (c and d) Concentration of adenylates in each section (moving average data) (error bars show root mean square deviations of the original data from the moving average data). (e and f) Adenylates expressed as percentages of the total. Symbols: ATP; *, ADP; *, AMP; O, total adenylates. E],
data therefore show considerable fluctuations. However, some of these fluctuations may also be due to variations in adenylate extraction. Some of the variability in the results has been removed by calculating a three-point moving average .of the data. The deviation of the original data from the moving average data was then calculated. Although this procedure is statistically more complicated, especially with regard to the calculation of the ECA deviation, these procedures enable any trends to be clearly visualized and any statistical significances in these trends to be estimated. All of the adenylates were present throughout the biofilm, suggesting that a considerable proportion of the bacteria were viable throughout the structure. The observation that total adenylates peak at some position across the biofilm which depends on biofilm age is interesting. The observed maximum is near the middle at 55 h but approaches the surface of the biofilm in the older samples. It is possible that this position corresponds to a peak in "healthy" microbial biomass. This suggestion has also been made by Kornegay and Andrews (16), who showed that once biofilm growth had reached a critical thickness of 70 pum, any increase did not lead to an increase in the rate of both dissolved oxygen and substrate removal. This indicated that the biofilm had a limited "active" layer. Differences in microbial biomass and in cell density have been observed in electron and light microscope studies and more recently by using sectioning techniques with cryoprotected biofilm to determine the number of viable cells through the biofilm
profile (15). The decrease in adenylates near the surface could be explained by irregular surface layers which produce incomplete sections. As a proportion of the total adenylates, the AMP level is higher in the lower depths of the biofilm and decreases toward the surface. There is an increase in the proportion of ATP toward the surface, while that of ADP remains fairly constant.
The change in ECA values from base to surface, although relatively small (roughly 0.2 units), could suggest that the deeper cells in the biofilm become diffusion limited for nutrient and/or oxygen. The predominantly low ECA value suggests that the cell population has a low energy status. It is unlikely that this is due to freezing the biofilm, since a low energy charge value for unfrozen, immediately extracted entire biofilm has previously been observed (15). Additionally, Dobbs and LaRock (7) reported that rapid freezing of marine sediment, in liquid nitrogen, did not lead to a significant loss of ATP relative to that found on immediate extraction of unfrozen samples. While the turnover in adenylates is rapid, it was considered that the sampling time of 1 to 3 min before freezing would have little effect on the proportions of adenylates through the biofilm. This was because the biofilm is effectively a buffered system by virtue of the relatively slow rates of solute exchange due to diffusion processes. Low ECA values have been noted before with spores, which may have ECA values of less than 0.1 (26). Chapman
VOL. 58, 1992
ADENYLATE MEASUREMENTS THROUGH BIOFILM
1633
TABLE 1. Additional data for replicate samples Replicate no. and sample time
Section no.
Replicate 1, 55 h
5 8 11 14 17 20 23 26
Replicate 2, 55 h
4.5 6.5 8.5 10.5 12.5 14.5 16.5 18.5 20.5
Replicate 1, 105 h
5 8 11 14 17 20 23 26 29 32 35
Replicate 2, 105 h
4.5 6.5 8.5 10.5 12.5 14.5 16.5 18.5 20.5 22.5 24.5 26.5 28.5 30.5 32.5
Total adenylates (pmol/section)
RMSDa (pmol/section)
% ATP
% ADP
% AMP
19.18 24.59 21.31 25.11 27.78 28.75 22.43 18.00
7.91 5.90 2.80 2.52 2.74 2.84 1.96 1.60
8.65 8.17 8.88 10.12 11.50 14.31 19.98 25.34
25.63 27.87 34.41 36.70 35.25 33.67 37.26 40.02
65.73 63.97 56.71 53.18 53.25 52.01 42.75 34.63
78.54 75.08 102.92 113.72 123.17 105.62 91.33 73.21 53.62
12.77 15.82 20.04 19.23 14.61 5.17 5.59 4.76 3.91
12.22 14.25 17.29 20.96 24.39 26.54 30.54 31.36 32.71
16.73 17.82 14.18 12.28 10.77
12.80 13.67 13.76
71.05 67.93 68.53 66.76 64.84 61.56 56.66 54.97 53.53
6.14 7.04 11.53 21.40 32.73 44.66 57.62 72.28 80.08 70.71 53.05
2.89 3.85 5.45 7.76 11.20 10.14 9.85 8.31 8.13 6.97 5.87
11.94 10.41 9.26 4.67 7.74 8.13 17.55 20.16 22.69 20.75 20.65
13.60 13.56 28.98 38.80 41.86 42.97 38.22 33.62 31.51 30.84 32.97
74.46 76.03 61.77 56.52 50.39 48.90 44.23 46.22 45.81 48.40 46.38
31.19 40.59 50.16 56.48 61.02 65.28 66.69 73.05 80.07 89.99 107.37 106.15 105.45 86.38 80.74
6.71 6.26 5.31 4.57 6.92 7.11 6.76 5.14 7.44 9.20 9.20 8.37 7.45 7.96 7.61
13.42 17.44 18.45 18.30
37.81 31.72 26.07 23.67 22.73 22.00 20.64 18.32 18.55 17.78 17.78 16.60
48.76 50.84 55.48 58.03 56.91 55.12 48.52 52.07 50.61 55.29 53.39 54.21 49.53 45.09 42.51
20.37 22.88 30.84 29.62 30.85 26.93 28.83 29.20 31.36 33.63 34.55
11.90
19.11
21.28 22.94
a RMSD, root mean square deviation of original data from moving average data.
and colleagues (4) showed that exhaustion of glucose from the growth medium caused the intracellular ECA of Eschenchia coli to decrease from about 0.8 to a value of 0.6 to 0.5. The lower value and viability were then maintained for 60 to 80 h. Low ATP values have also been reported for cells under starvation conditions which recovered rapidly upon replenishment of the nutrient supply (6). Whole biofilm of oral bacteria investigated by Wimpenny and colleagues (29) showed low steady-state film ECA values similar to those observed in this study. In their work, by using a similar perchloric acid extraction method, high ECA values could be induced in young biofilm exposed to high glucose concentrations. The biofilm generated in this study was produced by P. aeruginosa under the special circumstances of low nutrient concentration, and it is possible that there could be large
numbers of dead cells trapped within the biofilm, especially in the basal regions. It is clear, however, that the biofilm was growing and that it retained a large population of viable cells (ca. 2 x 108 per plug). Marshall (21) showed that starved bacteria adhering to surfaces were able to grow to normal size and to complete a number of cycles of cell division and concluded that if the limiting substrate were continually replenished, as in natural, flowing systems, rapid biofilm development would be expected. These observations raise the interesting question of what is normal for energy charge values in a structured, nutrientlimited habitat. There is some evidence for a wide range in ECA values in other structured ecosystems. Thus, Witzel (30) examined 356 water samples from several deep lakes in Germany. Witzel's results indicated that the ECA values ranged from 0.16 to 0.97. The highest ECA values were
APPL. ENVIRON. MICROBIOL.
KINNIMENT AND WIMPENNY
1634
(a)
(b)
55 h moving average data 1.0
0.9
e , u
e 0
105 h moving average data 1.0
0.8-
c
0.9 0.8
0.7-
4,
0.7
0.6-
u
0.6
0.5-
0
0.5
-
surface
0
S.
0.4-
2
C
0.3-
0
Vol. 58, No. 5
0099-2240/92/051629-07$02.00/0
Measurements of the Distribution of Adenylate Concentrations and Adenylate Energy Charge across Pseudomonas aeruginosa Biofilms S. L. KINNIMENT* AND J. W. T. WIMPENNY School of Pure and Applied Biology, University of Wales College Cardiff CFI 3TL, United Kingdom
Cardiff,
Received 16 December 1991/Accepted 27 February 1992
Adenine nucleotide pools and adenylate energy charge distributions were determined by using a laboratorygenerated quasi-steady-state Pseudomonas aeruginosa biofilm. The method used involved freezing and sectioning of the intact biofilm, followed by extraction and assay of the adenylates in the sectioned material. Results indicated an increase in adenylate energy charge of about 0.2 units from the bottom to the surface of the biofilm. However, energy charge values were generally low throughout the biofilm, reaching a maximum of only 0.6 units. Of the adenylates measured, AMP was the predominant nucleotide, especially in the deeper parts of the biofilm profile. Biofilm is the general term that applies to microbial communities forming coherent layers on solid surfaces. These structures are ubiquitous and can lead to serious economic loss to industry, besides posing health risks in the dental and medical spheres. These problems are exacerbated because biofilms are generally less sensitive to antimicrobial agents than are freely suspended organisms (22). As a spatially heterogeneous community, very little is known about the structure and physiology of biofilm. A number of experimental models which allow the investigation of biofilm throughout the growth cycle exist (10, 24, 28). The constant-depth film fermentor developed by Coombe and colleagues (5) and by Peters and Wimpenny (23) generates quasi-steady-state biofilm. This is defined on the basis that total protein and viable count remain constant over a period of time. Biofilm can be of any chosen thickness up to about 500 p,m. One indicator of the energetic status of living cells is the adenylate pool (1). Sensitive assays which are capable of measuring femtomole quantities of ATP exist (27). It was felt that this level of sensitivity would allow us to measure adenylates in 12-p,m-thick cryostat sections of biofilm. Adenine nucleotides, especially ATP, play a central role as intermediate carriers of chemical energy linking catabolism and biosynthesis. In vitro studies have revealed that the activities of certain enzymes are affected by the concentration of ATP; others are affected by ADP or AMP. The adenylate energy charge (ECA) is a measure of the effect of the ratios of adenosine phosphate concentrations on the rate of cellular metabolism (12) and is defined as follows: ECA = ([ATP] + 1/2[ADP])/([ATP] + [ADP] + [AMP]). The ECA iS a linear measure of the total amount of potential energy momentarily stored in the adenine nucleotide pool. It is a dimensionless number which ranges from 0 to 1 (2). Since there is claimed to be a positive correlation between the growth potential or functional capacity of a cell and its ECA (4), these measurements provide a framework for estimating metabolic potentials of naturally occurring microbial populations. This article describes a method for freezing and sectioning biofilm generated in the constant-depth film fermentor. By *
using this technique, adenylate pools and energy charge values have been determined as a function of depth through a Pseudomonas aeruginosa biofilm.
MATERIALS AND METHODS Organisms. A metalworking fluid isolate, identified as P. aeruginosa (14), was used to inoculate the fermentors. This microorganism was selected because it emerged as the predominant bacterium in the development of a stable biofilm derived from contaminated metalworking fluids. The microorganism was stored by freezing on glass beads at -70-C (11). Media. The amine-carboxylate medium used was a simplified simulated cutting fluid based on a mixture of fatty acids (predominantly oleic acid) and triethanolamine supplemented with mineral salts and vitamins. The medium contained, per liter of distilled water, an 0.085% (vol/vol) mixture of cutting fluid components consisting, by weight, of 60% triethanolamine and 40% fatty acids (of which 66% was oleic acid) and (in grams) the mineral salts FeSO4. 7H20 (1.5 x 10-2), MgSO4. 7H20 (0.2), ZnSO4- 7H20 (1.75 x 10-3),
MnSO4 * 4H20 (1 X
10-4), CUS04
5H20 (1 X
10-4),
NH4Cl (1.0), K2HPO4 3H20 (1.2), and CaCl2- 6H20 (1 x 10-2) and the vitamins inositol (1 x 10-2), calcium pantothenate (2.0 x 10-3), and biotin (1.0 x 10-5). For plate cultures, the same medium was used, except that the cutting fluid components were increased to 2% (vol/vol) and 2% (wt/vol) agar (Difco) was added. Cutting fluid components were supplied by Castrol Research Laboratories (Whitchurch Hill, Pangbourne, Reading, United Kingdom). Other chemicals were AnalaR grade (BDH) unless otherwise stated. Biofilm growth. Biofilm formation. The constant-depth film fermentor used by Kinniment and Wimpenny (14) was used to generate a reproducible biofilm. The constant-depth film fermentor consists of an enclosed rotating stainless steel disc in which are located 15 removable biofilm pans. Each biofilm pan is drilled with a central threaded hole, and around the edge of this are arranged six biofilm plugs. The latter are recessed to the selected depth up to 500 p.m by using a specially machined tool. Biofilm develops on the flat surface of the plug in the recessed space. The steel disc rotates beneath angled polytetrafluoroethylene (PTFE) scraper
Corresponding author. 1629
1630
KINNIMENT AND WIMPENNY
blades which remove excess growth, maintaining the biofilm at a constant depth. Medium drips onto the steel disc directly in front of one scraper blade. The latter distributes it over the biofilm pans, and any excess is wiped away. Samples of biofilm can be taken by replacing pans at selected time intervals by using two sterilizable stainless steel tools. A threaded extractor tool is used to remove and replace the pans, while a flat tamping tool is used to push replacement pans down flush within the turntable. The fermentor is heat sterilizable, the design allowing aseptic removal and replacement of biofilm pans. In addition, the system is operated under controlled gas phase and temperature conditions. The inoculum for the fermentor was prepared by aseptically flooding a 48-h-streaked amine-carboxylate plate of P. aeruginosa with 2 ml of sterile Tris-HCl buffer, pH 8.8. One milliliter of the homogenized mixture was then removed and resuspended in 4 ml of sterile buffer. This inoculum was allowed to recirculate through the fermentor in 500 ml of medium for 12 h. The main 10-liter batch of sterile medium was then aseptically connected to the fermentor. At the same time, the waste outlet port was connected to a large sterile waste receiver. The fermentor was operated with a medium flow rate of 1 ml min-1, an air flow rate of 250 ml min-1, a disc rotation speed of 3 rpm, and a constant temperature of 30°C. Sample pans were taken from the fermentor at selected time intervals. Each pan contained six PTFE sample plugs recessed to a depth of 300 p.m. The diameter of the plug surface was 4.75 mm, giving a surface area of approximately 17.72 mm2 and a biofilm volume, if the available space was completely full, of about 5.32 mm3. The fermentor was run to quasi-steady state before samples were removed for sectioning. The fermentor was closed down after 10 liters of medium had passed through the system. Estimation of total protein accumulation. Protein analysis based on a modified Lowry method was carried out (18). A sample pan was removed and four plugs were carefully pushed out of the pan, with the biofilm still intact on the surface. Protein on each plug was assayed separately by being boiled in 0.5 ml of 1 M NaOH for 5 min. These were each diluted fourfold in 1 M NaOH, and 0.5 ml (each) of the diluted mixtures was then used in the standard Folin procedure. Protein standards containing 5 to 150 p.g of protein (bovine albumin fraction five; BDH) were prepared each time the assay was performed. The absorbance values for the biofilm samples were then read against these standard curves. Viable count. The two remaining sample plugs from each pan were removed with flame-sterilized forceps. To disperse the biofilm, the plugs were transferred to 10 ml of sterile deflocculant solution in a universal container. The deflocculant consisted of 0.01% (wt/vol) Cirrasol (ICI Organics) in 0.01% (wt/vol) sodium pyrophosphate as described by Gayford and Richards (8). Approximately 250 2.5- to 3.5-mmdiameter glass beads (BDH) had been added to this solution. Sample plugs and biofilm were then agitated by vortex mixing for 90 s. After treatment, the sample was serially diluted in Tris-HCl buffer at pH 8.8 and 0.1 ml of each dilution was plated out in triplicate onto amine-carboxylate plates. The plates were incubated at 30°C for 24 to 48 h, and CFU were determined. Method for sectioning sampled biofilm. Before sectioning, a method for freezing and mounting the biofilm to a cryostat sample holder was developed, and it is described below. Agar (2% [wt/vol]) (Oxoid) plates were poured aseptically
APPL. ENVIRON. MICROBIOL.
to a depth of about 5 mm. Agar plugs were cut from these plates with a sterile no. 15-size cork borer. The plug of agar was removed from the petri dish with a sterile spatula, inverted, and placed on the upper surface of the agar remaining in the petri dish. A circle of lens tissue was placed on top of the agar plug, and a drop of water was placed on it to moisten it. With forceps, a cryostat sample holder was touched firmly to the surface of the damp lens tissue. The whole assembly was inverted and the sample holder, with the agar layer attached, was then lifted away from the surface of the agar plate and placed in a polystyrene container. Sample pans containing quasi-steady-state biofilm were removed from the fermentor at 55 and 105 h. The PTFE sample plugs were carefully removed from the sample pans. Sample plugs, biofilm upper surface down, were positioned between a pair of forceps which were then placed in the jaws of a micromanipulator. The whole assembly was placed directly above the sample holder, and 1 drop of water was quickly placed onto the surface of the agar. By using the micromanipulator, the PTFE sample plug with biofilm attached was lowered onto the drop of water so that the biofilm surface was just submerged. Liquid nitrogen was then poured into the polystyrene container to the depth of the agar plug, and the agar and biofilm were left to freeze. Once the biofilm was frozen, the forceps were removed from the micromanipulator and the PTFE sample plug was pulled away from the frozen biofilm. A modification of this method was used for one set of replicate samples; the main difference was that the sample holder was chilled to -37°C prior to being mounted onto the agar plug. The biofilm, previously frozen in liquid nitrogen, was then stuck onto the frozen agar surface with a drop of water, and once frozen the PTFE sample plug was removed. The time between taking samples of and freezing the biofilm was approximately 2 to 3 min for the first method and 1 to 2 min for the second. The sample holder was fixed to the microtome in a Starlet 2212 cryostat (Bright Instrument Company Ltd., Huntingdon, United Kingdom), and the knife blade was adjusted to cut 12-p.m, entire, horizontal biofilm sections. Sections were removed with a clean sterile microspatula. In the first method, sections were removed in threes; in the second method, the first section was discarded and subsequent sections were then removed in pairs. A total of six biofilm profiles were sectioned, three 55-h and three 105-h samples. Method of adenylate extraction of sectioned biofilm. The method described below was adapted from that used by Lundin and Thore (19) and Scourfield (25). Each set of two or three sections was transferred to 0.5 ml of 2.3 M perchloric acid containing 6.7 mM EDTA. The mixture was stirred constantly for 15 min on ice and centrifuged for 3 min in an Eppendorf bench top centrifuge at approximately 13,000 x g, and the supernatant was neutralized with 2 M KOH-0.5 M triethanolamine buffer. The mixture was left on ice for 5 min until all the potassium perchlorate had precipitated and then was centrifuged for a further 2 min. The supernatant was finally frozen at -70°C until needed. Adenylate and energy charge analyses of sectioned biofilm. ATP, ADP, and AMP were assayed in triplicate by standard methods (19, 20) by using a purified firefly lantern extract (Bio-Orbit Oy, Turku, Finland). An LKB 1251 computerinterfaced luminometer (Bio-Orbit Oy) was used for the bioluminescence assay. One hundred twenty-five microliters of the assay mixture was placed in a cuvette in the sample holder of the luminometer. Once the sample had been loaded
ADENYLATE MEASUREMENTS THROUGH BIOFILM
VOL. 58, 1992
(a)
suggests a doubling time of approximately 3.5 h
1-1 go
._: 0
I
10 0
25
_
I
50
75 100 Time (h)
125
75
125
150
17''5
lo 9
8.
X.
10
8
1 7
.0I
(P.max = 0.2
h-1) (Fig. lb). Analysis of adenylate concentration through the biofilm.
1000
(b)
1631
0I
52 10 6. 0
I
I
25
50
I
.
100
I
150
175
Time (h) FIG. 1. (a) Typical growth curve (each value expressed as miof protein per plug) of the biofilm. Error bars show standard deviations based on four sample plugs. (b) Viable count, expressed as CFU per plug (average for two sample plugs). crograms
into the detection chamber, 100 ,ul of ATP monitoring reagent (Bio-Orbit Oy) was rapidly injected, the sample and reagent were pulse-mixed, and the 10-s integral of light
emitted from each assay was recorded immediately. All assays were performed at 25°C. ATP was measured directly; ADP was determined after conversion to ATP by using pyruvate kinase (Sigma) and phospho(enol)pyruvate (Sigma); AMP was measured by being converted to ADP and then to ATP by using adenylate kinase (Sigma) together with the pyruvate kinase and phospho(enol)pyruvate. Internal standards were assayed to detect possible inhibition of the bioluminescence reaction, and an ATP calibration curve was also constructed.
The results of adenylate measurements showed random fluctuations (Fig. 2a and b); therefore, the data were smoothed by taking a three-point moving average. Although the first and last datum points are lost by using this approach, trends in adenylate levels and ECA values are easier to discern. Errors for the moving average adenylate concentrations and the ECA values were estimated by calculating the root mean square deviation of the original data from the moving average data for the former and then the ratios of random variables for the latter (13). An estimate of the probability of any significant difference in ECA between the base and surface points was calculated by comparing two samples, and then standard Student's t tables were consulted. Adenylate concentrations for two of the six biofilm profiles are shown in Fig. 2, as examples. For the remaining replicate samples, the total adenylates and the percentages of ATP, ADP, and AMP in relation to the total adenylates are presented in Table 1. The results for total adenylates and individual adenylates and the fraction of each expressed as a percentage of the total adenylates through the biofilm profile show similar trends for all six profiles (Fig. 2c to f and Table 1). Different biofilms were used for each profile, so some variation is to be expected. In the 55-h biofilms, concentrations of total adenylates, AMP, and ATP rose from the base and reached a maximum near the middle of the biofilm. By 105 h these peaks had moved toward the surface. ADP stays relatively constant, showing a small peak at the center of the 55-h biofilms and near the surface of the 105-h biofilms. All adenylate levels fall very near the surface of the biofilm in both sets of samples (Fig. 2c and d and Table 1). In the 55-h biofilms, the percentage of AMP in relation to the total adenylates drops toward the surface of the biofilm, that of ATP increases, and that of ADP stays relatively constant. For the 105-h biofilms, ATP and AMP follow a similar course but ADP either increases or decreases through the center before levelling off again near the surface. The dominant adenylate appeared to be AMP in all samples (Fig. 2e and f and Table 1). ATP levels were generally low. Values ranged from about 1 to 30 pmol per section. While the 55-h biofilms appeared to have reached a quasi-steady state, judging by the protein and CFU values, the number of sections taken from the 55-h samples was generally smaller than that of the 105-h samples. (Fig. 1 to 3). Examination of ECA through the biofilm. Energy charge values showed a consistent increase across both sets of biofilm from base to surface. While the range was not large, it represented a value of about 0.2, from values of 0.22 to 0.28 near the base to 0.4 to 0.45 near the surface of the 55-h biofilms (Fig. 3a) and 0.17 to 0.35 near the base to 0.37 to 0.6 at the surface of the 105-h biofilms (Fig. 3b). The probability of a significant difference between the points at the biofilm base and the surface ranged from 70 to 99.9%, with an average value of 90%.
RESULTS
Biofilm growth. The P. aeruginosa biofilm was grown in amine-carboxylate medium for a period of 150 h. Quasisteady-state conditions were reached after about 45 h. Quasi-steady-state protein values plateaued at about 250 jig per plug (Fig. la). The viable count was about 2 x 108 CFU per plug. The slope of the log of viable count versus time
DISCUSSION A number of different sectioning procedures were tested before the methods described in this article were selected. It was not always easy to produce complete sections, and occasionally material was lost during the transfer of sections from the microtome blade to the sample container. The raw
KINNIMENT AND WIMPENNY
1632
APPL. ENVIRON. MICROBIOL.
(c)
(a)
(e)
55 h moving average data
55 h orginal data
1
au
C)
* .2 5)
5).r
70
base
60
4+
50 Cu 40 I 30 0 20 10
2,
55 h moving average data -
surface
+ =
-
0I 0
(b)
10 20 Section number
30
0
10
20 Section number
30
0
(D
105 h original data
10 20 Section number
30
105 h moving average data
120
70- base
rA100. n080 0 ° 80
0-
4
44
c -%
50-
1
cAf4 10 'e
surface
40V 30o 2010 -
53
0 Section number
Section number
I
n
10 20 Section number
30
FIG. 2. Graphical examples of adenylate concentrations through the biofilm. (a and b) Concentration of adenylates in each section (original data) (error bars show standard deviations of triplicate samples). (c and d) Concentration of adenylates in each section (moving average data) (error bars show root mean square deviations of the original data from the moving average data). (e and f) Adenylates expressed as percentages of the total. Symbols: ATP; *, ADP; *, AMP; O, total adenylates. E],
data therefore show considerable fluctuations. However, some of these fluctuations may also be due to variations in adenylate extraction. Some of the variability in the results has been removed by calculating a three-point moving average .of the data. The deviation of the original data from the moving average data was then calculated. Although this procedure is statistically more complicated, especially with regard to the calculation of the ECA deviation, these procedures enable any trends to be clearly visualized and any statistical significances in these trends to be estimated. All of the adenylates were present throughout the biofilm, suggesting that a considerable proportion of the bacteria were viable throughout the structure. The observation that total adenylates peak at some position across the biofilm which depends on biofilm age is interesting. The observed maximum is near the middle at 55 h but approaches the surface of the biofilm in the older samples. It is possible that this position corresponds to a peak in "healthy" microbial biomass. This suggestion has also been made by Kornegay and Andrews (16), who showed that once biofilm growth had reached a critical thickness of 70 pum, any increase did not lead to an increase in the rate of both dissolved oxygen and substrate removal. This indicated that the biofilm had a limited "active" layer. Differences in microbial biomass and in cell density have been observed in electron and light microscope studies and more recently by using sectioning techniques with cryoprotected biofilm to determine the number of viable cells through the biofilm
profile (15). The decrease in adenylates near the surface could be explained by irregular surface layers which produce incomplete sections. As a proportion of the total adenylates, the AMP level is higher in the lower depths of the biofilm and decreases toward the surface. There is an increase in the proportion of ATP toward the surface, while that of ADP remains fairly constant.
The change in ECA values from base to surface, although relatively small (roughly 0.2 units), could suggest that the deeper cells in the biofilm become diffusion limited for nutrient and/or oxygen. The predominantly low ECA value suggests that the cell population has a low energy status. It is unlikely that this is due to freezing the biofilm, since a low energy charge value for unfrozen, immediately extracted entire biofilm has previously been observed (15). Additionally, Dobbs and LaRock (7) reported that rapid freezing of marine sediment, in liquid nitrogen, did not lead to a significant loss of ATP relative to that found on immediate extraction of unfrozen samples. While the turnover in adenylates is rapid, it was considered that the sampling time of 1 to 3 min before freezing would have little effect on the proportions of adenylates through the biofilm. This was because the biofilm is effectively a buffered system by virtue of the relatively slow rates of solute exchange due to diffusion processes. Low ECA values have been noted before with spores, which may have ECA values of less than 0.1 (26). Chapman
VOL. 58, 1992
ADENYLATE MEASUREMENTS THROUGH BIOFILM
1633
TABLE 1. Additional data for replicate samples Replicate no. and sample time
Section no.
Replicate 1, 55 h
5 8 11 14 17 20 23 26
Replicate 2, 55 h
4.5 6.5 8.5 10.5 12.5 14.5 16.5 18.5 20.5
Replicate 1, 105 h
5 8 11 14 17 20 23 26 29 32 35
Replicate 2, 105 h
4.5 6.5 8.5 10.5 12.5 14.5 16.5 18.5 20.5 22.5 24.5 26.5 28.5 30.5 32.5
Total adenylates (pmol/section)
RMSDa (pmol/section)
% ATP
% ADP
% AMP
19.18 24.59 21.31 25.11 27.78 28.75 22.43 18.00
7.91 5.90 2.80 2.52 2.74 2.84 1.96 1.60
8.65 8.17 8.88 10.12 11.50 14.31 19.98 25.34
25.63 27.87 34.41 36.70 35.25 33.67 37.26 40.02
65.73 63.97 56.71 53.18 53.25 52.01 42.75 34.63
78.54 75.08 102.92 113.72 123.17 105.62 91.33 73.21 53.62
12.77 15.82 20.04 19.23 14.61 5.17 5.59 4.76 3.91
12.22 14.25 17.29 20.96 24.39 26.54 30.54 31.36 32.71
16.73 17.82 14.18 12.28 10.77
12.80 13.67 13.76
71.05 67.93 68.53 66.76 64.84 61.56 56.66 54.97 53.53
6.14 7.04 11.53 21.40 32.73 44.66 57.62 72.28 80.08 70.71 53.05
2.89 3.85 5.45 7.76 11.20 10.14 9.85 8.31 8.13 6.97 5.87
11.94 10.41 9.26 4.67 7.74 8.13 17.55 20.16 22.69 20.75 20.65
13.60 13.56 28.98 38.80 41.86 42.97 38.22 33.62 31.51 30.84 32.97
74.46 76.03 61.77 56.52 50.39 48.90 44.23 46.22 45.81 48.40 46.38
31.19 40.59 50.16 56.48 61.02 65.28 66.69 73.05 80.07 89.99 107.37 106.15 105.45 86.38 80.74
6.71 6.26 5.31 4.57 6.92 7.11 6.76 5.14 7.44 9.20 9.20 8.37 7.45 7.96 7.61
13.42 17.44 18.45 18.30
37.81 31.72 26.07 23.67 22.73 22.00 20.64 18.32 18.55 17.78 17.78 16.60
48.76 50.84 55.48 58.03 56.91 55.12 48.52 52.07 50.61 55.29 53.39 54.21 49.53 45.09 42.51
20.37 22.88 30.84 29.62 30.85 26.93 28.83 29.20 31.36 33.63 34.55
11.90
19.11
21.28 22.94
a RMSD, root mean square deviation of original data from moving average data.
and colleagues (4) showed that exhaustion of glucose from the growth medium caused the intracellular ECA of Eschenchia coli to decrease from about 0.8 to a value of 0.6 to 0.5. The lower value and viability were then maintained for 60 to 80 h. Low ATP values have also been reported for cells under starvation conditions which recovered rapidly upon replenishment of the nutrient supply (6). Whole biofilm of oral bacteria investigated by Wimpenny and colleagues (29) showed low steady-state film ECA values similar to those observed in this study. In their work, by using a similar perchloric acid extraction method, high ECA values could be induced in young biofilm exposed to high glucose concentrations. The biofilm generated in this study was produced by P. aeruginosa under the special circumstances of low nutrient concentration, and it is possible that there could be large
numbers of dead cells trapped within the biofilm, especially in the basal regions. It is clear, however, that the biofilm was growing and that it retained a large population of viable cells (ca. 2 x 108 per plug). Marshall (21) showed that starved bacteria adhering to surfaces were able to grow to normal size and to complete a number of cycles of cell division and concluded that if the limiting substrate were continually replenished, as in natural, flowing systems, rapid biofilm development would be expected. These observations raise the interesting question of what is normal for energy charge values in a structured, nutrientlimited habitat. There is some evidence for a wide range in ECA values in other structured ecosystems. Thus, Witzel (30) examined 356 water samples from several deep lakes in Germany. Witzel's results indicated that the ECA values ranged from 0.16 to 0.97. The highest ECA values were
APPL. ENVIRON. MICROBIOL.
KINNIMENT AND WIMPENNY
1634
(a)
(b)
55 h moving average data 1.0
0.9
e , u
e 0
105 h moving average data 1.0
0.8-
c
0.9 0.8
0.7-
4,
0.7
0.6-
u
0.6
0.5-
0
0.5
-
surface
0
S.
0.4-
2
C
0.3-
0
