APPLIED AND ENVIRONMENTAL MICROBIOLOGY, June 1982, p. 1227-1237 Vol. 43, No. 6 0099-2240/82/061227-11$02.00/0
Attached and Free-Floating Bacteria in a Diverse Selection of Water Bodies COLIN R. BELL AND LAWRENCE J. ALBRIGHT* Department ofBiological Sciences, Simon Fraser University, Burnaby, British Columbia, Canada VSA IS6 Received 21 September 1981/Accepted 3 February 1982
The contribution of attached and free-floating bacteria to the bacterial numbers and heterotrophic uptake in 44 diverse aquatic environments was studied. A factor analysis reduced the variability of the raw data base to three major factors explaining 53.6% of total variance. These factors were (i) salinity, (ii) heterotro- phic uptake, and (iii) particulate load. A cluster analysis categorized the 44 habitats into five distinct environmental types based on these three factors. There was no significant pattern in the distribution of attached versus free-floating bacteria when assessed by epifluorescent microscopy. However the contribution of attached bacteria to the uptake of an amino acids mix was reduced in marine waters. Heavy particulate loads resulted in an increased percentage uptake of amino acids and glucose from the attached bacteria. Uptake response was found to be substrate specific especially in oliogotrophic freshwater. Amino acid uptake was more associated with the attached fraction, whereas glucose uptake was mediated more by the free-floating fraction.
The association of bacteria with particulate from the depths listed in Table 1 with a 7-liter Van material is of considerable importance in aquatic Dorn bottle. All samples were processed immediately environments. Bacteria are one of the major after collection. agents responsible for the alteration of particu- Physicochemical determinations. Temperature was measured with a simple mercury thermometer. pH and late matter through processes such as decay and salinity and conductivity were determined with a Fish- flocculation (16-18). They are able to convert er model 107 portable pH meter and a Yellow Springs dissolved organic matter to particulate organic Instruments model S-C-T 33 salinometer, respective- matter (2), and their ultimate fate in the water ly. The total volume and modal size of particulates will be greatly influenced by the material to were counted later in the laboratory with a Coulter which they are attached (20). Counter TA II with 280- and 100-rm-diameter aper- Most authorities agree that in the open ocean tures. This covered particles in the diameter size range planktonic bacteria are unattached and free of 1.26 to 80.64 ,um. Seston was determined as an In freshwater environments estimate of the organic particulates in the waters. A floating (6, 22, 24). known volume ofwater was filtered through a predried the degree of attachment varies considerably (4, and tared Whatman GFF filter which was fixed by 5, 7, 8, 14, 19). However evidence is accumulat- placing in a petri plate with a pad soaked in Formalin. ing in studies on rivers that as the freshwater The filter was later ashed in a furnace at 500°C for 24 h. flows into the sea the degree of attachment The weight of material burned off was reported as decreases with increasing salinity (9, 25). seston. Particulate organic carbon (POC) was defined We observed a similar decrease in numbers as the organic carbon retained on a GFF filter previ- and activities of attached bacteria in the Fraser ously baked for 24 h at 500°C. The filtrate was consid- River Estuary as it entered the Strait of Georgia ered to be dissolved organic carbon (DOC). Both POC To assess how universal the and DOC samples were fixed in the field by freezing in (3). relationship dry ice and were kept frozen until ready for analysis. between attachment and salinity was and also to POC was determined with a Perkin Elmer Elemental explore other influences, we embarked on a Analyser 240. DOC was analyzed on a Beckman study of 44 aquatic environments, 15 of which Tocamaster model 915-B. were coastal marine. The results of this study Information on the location and elevation of each are reported herein. sample site was taken from topographical maps. Two location parameters, called X coordinate and Ycoordi- MATERIALS AND METHODS nate, were entered into the statistical analyses to Collection of water samples. A total of 44 water detect any influences attributable to the proximity of samples were collected over the period 26 May 1980 to the urban and industrial centre of Vancouver. The X 25 November 1980. The majority of these samples (40) coordinate measures minutes of longitude west of were taken from the surface water in sterile Nalgene sample no. 32, Lost Lagoon. The Y coordinate mea- containers. The remaining four samples were collected sures minutes of latitude north of sample no. 32. Lost
1227 1228 BELL AND ALBRIGHT APPL. ENVIRON. MICROBIOL. TABLE 1. Site number and location Site no. Location Longitude Latitude 1 Porlier Pass 123035'W 49001'N 2 Active Pass 123019'W 48052'N 3 Nanaimo River Estuary 123054'W 48008'N 4 Howe Sound (250 m) 123015'W 49005'N 7 Howe Sound (8 m) 123015'W 49005'N 8 Indian River Estuary 122053'W 49029'N 10 Indian Arm (200 m) 122052'W 49023'N 11 S.F.U. Reflecting Pool 122055'W 49017'N 12 Fraser River 122055'W 49012'N 13 Burnaby Lake 122057'W 49015'N 14 Buntzen Lake 122052'W 49020'N 15 Chadsey Lake 122009'W 490 TN 16 Marion Lake 123010'W 49033'N 17 Cheakamus Lake 122056'W 50001'N 18 Lindeman Lake 121°27'W 49007'N 19 Levette Lake 123011'W 49050'N 20 Pacific Ocean 125040'W 49043'N 21 Pacific Ocean (150 m) 125040'W 49043'N 22 Alberni Inlet 124048'W 49009'N 23 Trevor Channel 125009'W 48051'N 24 Goldie Lake 122056'W 49022'N 25 Henriette Lake 123020'W 49041 'N 26 First Lake 123°11'W 49023'N 27 Wedgemount Lake 122049'W 50010'N 28 Lafarge Lake 122046'W 49017'N 29 Killarney Lake 123021 'W 49024'N 30 Greendrop Lake 121026'W 49009'N 31 Mill Lake 122019'W 49003'N 32 Lost Lagoon 123008'W 49018'N 33 Garibaldi Lake 123001'W 49055'N 34 Hicks Lake 121°42'W 49020'N 35 Pitt Lake 122035'W 49021 'N 36 Serpentine River 122045'W 49008'N 37 Squamish River 123016'W 49053'N 38 Chilliwack River 121058'W 49006'N 39 Bute Inlet 124050'W 50053'N 41 Homathko River 124052'W 50057'N 42 Calm Channel 125006'W 50021 'N 43 Robert Burnaby Park Creek 122056'W 49014'N 44 Post Creek 121°29'W 49006'N 45 Steveston Drainage Ditch 123005'W 49008'N 46 Fraser River Estuary 123008'W 49°06'N 47 Georgia Strait 123030'W 49009'N 50 Sewage Effluent (Annacis Island) 122058'W 49°10'N
Lagoon was chosen as a convenient reference point as and 97% of all particulates in the estuary were retained it is situated in downtown Vancouver. by this filter, yet in samples which had a predominance Biomass determinations. Chlorophyll a was deter- of free-floating bacteria, 98% of the bacteria passed mined spectrophotometrically after acetone extraction through the filter. Samples were fixed in the field with with Millipore 0.8-,um pore size filters (23). The filters 2% (vol/vol) Formalin. were removed from the field wrapped in foil and Heterotrophic uptake. The turnover time of glucose analyzed within a few days. Bacterial numbers were and an amino acid mix was calculated by the method estimated by the acridine orange direct count (AODC) of Azam and Holm-Hansen (1). D-[6-3H]glucose (spe- method of Hobbie et al. (10). Bacteria attached to cific activity, 30 Ci/mmol; New England Nuclear particles, which included bacterial aggregates and bac- Corp.) and a mixture of 15 tritiated L-amino acids teria attached to plankton, were distinguished from (New England Nuclear Corp.; code no. NET-250) free-floating bacteria after removal of the bulk of the were diluted to a working concentration of 10 ,uCi/ml. particles by gently filtering 10 ml of water sample The relative uptake of each amino acid within the through Nuclepore 1.0-,m pore size filters. This pro- mixture could not easily be determined, so calcula- cedure has been tested extensively in the varying tions to specific concentrations of amino acids and conditions of the Fraser River Estuary (3). Between 60 glucose were not attempted. The water sample was VOL. 43, 1982 BACTERIA IN A SELECTION OF WATER BODIES 1229
SITE No. 01 39 02 go 23 CLUSTER A 03 33 08 20 hUIllQ 07 47 21 .XX1K 9, CLUST~ER B 42 46 10 +X+xs ".. 05 14 38 -+-. .+. XX+lw I 18 .- - ++-XX o 33 _,+ +++IXt so 30 -+ -..X+ XX- XX onel 25 27 _a+-. ___,~.w +_ seeH is 44 -X+--j. * ++- X+ 00|lm -X*- *x ++, ,_ 19 #-x+- *I[.+ +,w 0Z0+gme "I CLUST'ER C 16 +X+--X-X XX333 stee lose 43 +X+--X-XXXXNX_... ..- --.+- 30||l133le 37 35 Dwellsl 34 _ X+._,-XXXX + X 33311l 24 _+ _,,--X-GINN mx:nsmemls 26 , _ , , _,- -- X 17 -+.-.-.-. O.NX 3333 15 **...* 3333 45 A--MTIfa 32 11111111 CLUSTER D 36 4*l+xx..-.--+.-.-t!..X...xxxx x . +-x~+XXXXUUNXx so 41 XX++.-XXXXX I+ !31X+.XI+ XX3333Name 12 XIXX+XXXXXY1I* +i1X IXXN133333313 22 * xUIUN..... * - --.e -+*! 31 + IXXXX..--. +. . 1+--*1131310331D D+N1. 3J1 3e 13 -+"xxx01,00 . . .0 .--X +XI 1+X.--.I + II CLUSTER E 28 + -+ XJ a . -. *..XXU1XX11X1.-- Xl- 29 ++Xxi _ -. .t++1XX+lN+X.-- Xi'- 1 1 .--+XX . . --+.--X X-+ . 50 d.o3
FIG. 1. Distance matrix of sites arranged in clusters. The distances are represented in shaded form according -, to the following scheme: 0, <0.953; 0, 0.953 to 1.487; U, 1.487 to 2.002; x, 2.002 to 2.268; +, 2.268 to 2.535; 2.535 to 2.802; -, 2.802 to 3.279; open space, >3.279. 1230 BELL AND ALBRIGHT APPL. ENVIRON. MICROBIOL.
prepared by pouring four 100-ml volumes, previously o en " 00 £o -aq. o V- WI,{ filtered through a 300 Fm Nitex mesh, into four 0 eq en -- beakers 00 C4 r- N- % rwe~ I supported in the water body with a small U styrofoam raft to keep the samples at in situ tempera- ture. Two beakers were treated as controls by adding 0s , "uNN 2% (vol/vol) Formalin. Glucose solution (200 ,ul) was 0c) OEXs u then added to a test and a control beaker, similarly 00 --HP ,, t ") ON%- 0% 00 _ tt with the amino acid solution. The samples were then a C) 1%;r4 r-:~ incubated in the dark for up to 2.5 h. Samples (10 ml) were removed every 30 min and filtered by a proce- 4._ 0 "t ON en 00 WI, e) e in'*) dure previously described (3) to distinguish uptake rm0 5- rf,,q. mediated by attached bacteria from that mediated by .) lzuE. free-floating bacteria. Briefly, the samples were first filtered through 1.0-,um pore size Nuclepore filters, co -a O c en " Ch and then the filtrate was passed through 0.22-,um pore 0 0 00 00 oO'r F en " size Millipore filters. The radioactivity on both filters ao 8 tn was later counted on a Beckman LS 8000 scintillation o,,C1N0%r)_, counter, correcting for quench with the external stan- 0rCU C) 00 ent dards ratio method. tnmQ 0 oo Microautoradiography. Numbers of actively metab- 1.0- e- l --l bacteria - olizing were determined ES 1.It (NAB) by the tech- C) nique of Hoppe (13) with the following modifications. Samples of water (1 ml) were filtered through 300-p.m Nitex mesh and then placed in four test tubes and held $-4 C1) - I s- n at in situ temperature in the raft. Two tubes were *tEEt s 44 r-- r-- ;s14 ^E O treated with 2% (vol/vol) Formalin as controls. Tritiat- e o0X ed glucose and amino acid solutions (100 X _ ,uCi/ml) were -a then added, 50 ,ul to a tube, and incubated in the dark C) no - e- r- a, C) In o i for 3 h. At the end of this period, the two test samples CU I1 0;cl were fixed with 2% (vol/vol) Formalin, and then all 5- to 000% 0%0000 four tubes were brought to a final volume of 10 ml with C 0%N 00r-C la sterile isotonic water. Samples were filtered through C) 0 4.0 1.0-p.m pore size 5't o 0 Nuclepore filters and washed five '0 times with 10 ml of isotonic water. The filtrate was CU -4 00 00 0 saved and then filtered through 0.2-p.m pore size 0z 0 m0 "m Nuclepore filters and washed five times. After air drying, the filters were fixed to microscope slides with .1O 0 4.) rm EX C1) V- 0 00Z0 double-sided sticky tape and returned to the labora- 0s4X v., < 4 c~' x o et) tory. The slides were dipped in Kodak NTB-2 emul- VS sion and stored at 5C for 7 days in black plastic boxes. F-N The microautoradiograms were developed in Kodak F0 K 0% r 00 N= 0 E- r-* -4 D-19 for 5 min and then hardened in Anti-Scratch e4 m o00 00 Hardener (Edwal Scientific Products Corp., Chicago, 0 00 Nm Ill.) for 1 min before fixing for 10 min in Edwal Quick 0. 0 Fix. The slides were finally washed for 1 h in tap '0 water. C.) vo en o o Silver grains were counted by using a Zeiss V%0e6 00C 0 Standard WL microscope under bright field illumina- :3 -m X~rV - o m tion at 1,250x magnification. '0 Statistical analyses. All analyses of the data were 1- performed with the computer packages BMDP Bio- 0.C8 lIi en medical Computer Program P-series 1979 or Statistical ,- *. Package for the Social Sciences, 2nd ed. The normali- .0 v C o00 00 xe la - rn ty of all variables was checked with BMDP5D. It was necessary to perform a log transformation on the following variables to enhance normality and homoge- .0c=o V- 1%0 "' 0=00 neity of variance: conductivity, salinity, particle vol- U3 ume, elevation, seston, POC, DOC, chlorophyll a, AODC, NAB, and turnover time 0 estimates. The factor 8 analysis was performed on all independent variables * *. by using BMDP4M. The correlation matrix was fac- tored by principal component analysis with varimax r- C 00 rotation offactors. Factor scores were then computed evs N6 el; I for all 44 sites, and the scores on the first three factors N n000 were used to construct the pin diagram in Fig. 1. The cluster analysis of sites was obtained by using BMDP2M with factors 1, 2, and 3 as criteria. Clusters VOL. 43, 1982 BACTERIA IN A SELECTION OF WATER BODIES 1231 were formed by average linkage based on the Euclid- ean distance between sites (i.e., the square root of the sum of squares of the differences between the values of variables for each two sites). The arrangement of CD data by cluster (see Table 6) was performed by 0.9 00 p CD to BMDP7D with a one-way analysis of variance. The 0 three-way analysis of variance was executed by Statis- CDXp5 0%A -.4 0%> -0)E - -.4 Description of sites. The 44 water bodies sam- 1. X X pled comprised a random selection of marine, CD v.0v estuarine, lake, river, and creek environments in f9S~~~~~L0 0 the southwest corner of British Columbia as - CD 0- 00 -.4 described in Table 1. Table 2 indicates the l+ 00 OD * * 03. diversity of the sites with the more obvious . 3 1111 * 1+ microbiological parameters showing an exten- I+ 10 tI sive range of values, e.g., temperature, 8 to C 5 r 25°C; salinity, 0 to 30%; POC and DOC, 2.0 to 59.2 chlorophyll a, 0.1 to 39.7 mg/liter; mg/m3; 0 and total bacterial counts, 2.0 x 104 to 6.5 x 106 + 00 i04_. per ml. The total population investigated there- 1+ o l+ I+ M. a the C_c fore represented broad selection of niches (A of heterotrophic bacteria. - o Attached and free-floating bacterial activity. *o CL -). g %O C 3 The aspect of these environments of interest in 0 this study was the contribution of free-floating 0 0~0. and attached bacteria to biomass and heterotro- CD phic uptake. Table 3 shows the data averaged for some pertinent variables which indicated that, CD 1T+ %O--I+ I+ 0)C 3E-so 3 over the whole data base, the heterotrophic cr A 0 A F- CDM uptake (turnover time) of glucose and amino CD0 Is (A acids was greater for the free-floating portion - 1.o-i than for the attached portion. The percentage of M . uptake attributed to the attached flora for amino - z acids (46%) was also significantly greater than .-, CO) that for glucose (31%). However, the greater -£ o oS5c activity of the free-floating fraction was not CD paralleled by an increase in the number of free- 0 CD 1+ X I+ floating bacteria as estimated by AODC; the numbers in both fractions were comparable. 5CD Data on the number of active bacteria and O turnover time per active cell demonstrated a > x x different response depending on the radioactive substrate. The bacterial population actively me- > XaX amino acids had greater 00 tabolizing significantly 0U '. counts in the free-floating fraction compared 0C with the attached fraction. When glucose was t0. used as the substrate, the NAB were similar, whereas the turnover time per active cell was significantly lower in the free-floating fraction. - CD Factor analysis. The breakdown of the data in z ID Table 3 into counts of sites where the free- C) floating fraction was greater than the attached 0 co ITC- fraction of a given variable only showed a 1.41 marked difference in the turnover time of glu- 3Ei t-. cose, where 34 sites showed higher activity in vA. the free floaters. To explore the data base more 1232 BELL AND ALBRIGHT APPL. ENVIRON. MICROBIOL. TABLE 4. Matrix of rotated factor loadings Turnover time Factor Conduc Salinity Elevation pH Seston Particle X coor- Glucose Amino acids Free Attached Free Attached 1 0.97 0.97 -0.88 0.82 0.73 -0.69 0.65 0.49 2 0.87 0.70 0.81 0.48 3 0.52 0.54 -0.48 -0.34 4 -0.28 -0.41 5 0.55 6 Communality 0.95 0.96 0.87 0.73 0.84 0.79 0.79 0.78 0.73 0.77 0.75 a Total percent variance explained. thoroughly a factor analysis was performed (Ta- ture. All these variables are characteristic of the ble 4) which produced six factors explaining ocean environment as is the strong west influ- 73.7% of the variance. ence of the X coordinate. This first factor, Factor 1 received high loadings on salinity, responsible for 23.2% of total variance, was conductivity, pH, and seston with negative load- therefore interpreted as marine influences. ings on elevation, particle volume, and tempera- Factor 2 was characterized by high positive
3
45 2
26
41 32 34 36 Y 12 1 18 24 20 -
t t1 2t42 10 I-476 13- 'Tii~2L
28 -2
-3 -3 /~~~~~~2
3 _-% _ FIG. 2. Three-dimensional pin diagram of sites 1 to 50 by factor scores for factors 1, 2, and 3. Symbols: 0, cluster A; *, cluster B; A, cluster C; O, cluster D; *, cluster E. VOL. 43, 1982 BACTERIA IN A SELECTION OF WATER BODIES 1233 TABLE 4-Continued NAB AODCPatce%O Chloro- DOC POC Y coor- | GlucoseAmino as aice Te Eigen variance GlucosAminoacphylla Free Attached dinate mode value explained Free Attached Free Attached by factor -0.34 -0.43 5.34 23.2 -0.55 -0.29 -0.54 -0.32 -0.45 0.40 -0.49 3.63 15.8 0.36 0.39 0.43 0.81 0.80 0.66 -0.50 3.37 14.6 0.64 -0.37 0.56 0.44 1.69 7.3 -0.48 0.89 1.53 6.7 0.34 -0.34 0.34 0.89 0.34 1.41 6.1 0.63 0.53 0.54 0.69 0.65 0.71 0.78 0.65 0.53 0.86 0.87 0.56 73.7a loadings on turnover with negative loadings on sites were again all marine, but with high turn- chlorophyll a, NAB, temperature, and AODC. over times. The largest cluster formed was clus- This signified a factor of weak heterotrophic ter C, comprising 18 sites which include sites 14, activity with correspondingly low microbial flo- 38, 18, 33, 30, 25, 27, 44, 19, 16, 43, 37, 35, 34, ra. 24, 26, 17, and 15. These were all freshwater Factor 3, in contrast to factor 2, had high sites (negative factor 1) with high turnover times positive loadings on AODC, NAB, and chloro- (positive factor 2). Sites 45, 32, 36, 41, and 12 phyll a together with seston, particle volume, constituted cluster D, which was the only group POC, and DOC. This indicated a factor con- to receive a significant positive loading on factor cerned with high productivity, especially high 3. These sites were therefore characterized by particulate load. having heavy particulate load, freshwater (nega- Factors 4, 5, and 6 were not as significant as tive factor 1), and a range ofturnover times. The the first three factors; together they only ac- remaining cluster, cluster E, contained sites 22, counted for 20.1% of total variance and were far 31, 13, 28, 29, and 11. They were predominantly harder to interpret. Factor 4, because of the freshwater habitats with low particulate loads, positive loadings on the two NAB (free) varia- but very low turnover times (negative factor 2). bles, was believed to reflect some influence of Attached and free-floating bacterial activity by the free-floating active bacteria. Factor 5 was duster. Examination of the data arranged into obviously a location influence with a strong five clusters (Table 5) indicated that the percent northwest bias, whereas factor 6 alluded to the activity attached for both glucose and amino larger particulate loads. acids had significant differences between clus- Cluster analysis. The first three factors ap- ters. The order of clusters with increasing activi- peared very relevant to the question of attached ty from the attached fraction for amino acids versus free-floating bacteria and associated het- was A < B < E < C < D, whereas the order for erotrophic uptake. These three factors account- glucose was A < B < C < E < D. The total ed for 53.6% of the total variance in the popula- turnover times for both substrates likewise tion and were used as the three criteria to cluster showed significant differences. The ranking for the sites. The distance matrix (Fig. 1) contained amino acids and glucose was identical: E < D < five discrete clusters with all elements having A < C < B. The other variable with significant very small distances between each other difference of variance was the total epifluores- (-2.002 units). Site 50, the primary effluent from cent counts. Clusters E and D were most notable Annacis Island sewage plant, was the only ex- in possessing far higher bacterial counts than ception and was so dissimilar from the rest of the clusters A, B, and C. population that it was excluded. Figure 2 shows Although the turnover time per active cell the sites drawn in three-dimensional factor exhibited no significant differences between space with their respective clusters indicated. clusters, inspection of the number of cases with- Cluster A contained sites 1, 39, 2, 23, and 8, all in clusters where the free-floating fraction was situated in the quadrants with positive factor 1 greater than the attached, and vice versa, did loadings and negative factor 2 and 3 loadings. show interesting patterns. All of the turnover This group therefore tended to have a strong time estimates for clusters A and B showed that marine component with low turnover times and in the majority of sites the attached portion were low particulate load. Cluster B contained sites larger than in the free-floating portion. This 20, 7, 47, 21, 42, 46, 10, and 5, which all implied that the activity of the free floaters was displayed positive factor 1 and factor 2, with greater in most cases. The converse applied to factor 3 near zero. This implies that cluster B cluster D, where more than half the sites had 1234 BELL AND ALBRIGHT APPL. ENVIRON. MICROBIOL. greater activities in the attached fraction. Clus- 00 ter a r-1 N 00 -" C showed division between attached and 13 bacteria on o It - ') 00" 0 free-floating dependent the sub- 00c +1 < +1 < +1 < + strate. Turnover times for amino acids were S. +i <1 r for the 0 o~~~~- 0 longer free-floating population in approx- 0) o e0 00 r". 8 a 4 imately 75% of the sites comprising the cluster. 0. The opposite situation occurred with glucose, 0) where the greater activity was associated with the free-floating bacteria. I..co 0) 10 0 O en Factors influencing percent activity attached. 0% 0 o0 (_4 N '40 41 N^ The three-way analysis of variance tales (Tables 0 6 and 7) demonstrated that factor 3 had a signifi- 0) E_ cant influence on the percent activity attached 0) _0 with both amino acids and glucose. The contri- 0) C o Hb f cn bution of the attached population was greater 0 r- ,~ C 3U when factor 3 had positive loadings signifying o' e. e1H Not high particle load. The uptake of amino acids qQ0 0 0 o) between the free-floating and attached flora was C) CU also influenced by factor 1 (Table 6). A lower (4 '0 percent activity attached was recorded when , - _ -o factor 1 was positive, i.e., when the environ- Ir- 'ot en +1 <+1 00+1 <+1 00 ments showed a strong marine character. A 000 vC closer analysis of this relationship between per- e) cent activity attached and factor 1/factor 3 ratio 0 .r 4-- is shown in Table 8. The grouping offactor 1 into S.0;^ 0 + n Mean Multiple Of Factor Category score R2squaresSource of variation Sum df F Significance 1 Negative 28 56.2 0.23 Positive 16 28.2 2 Negative 17 43.8 0.004 Positive 27 47.4 3 Negative 30 37.1 0.22 Positive 14 65.2 Main effects 12,795 3 7.69 <0.001 Factor 1 5,258 1 9.48 0.004 Factor 2 19 1 0.03 0.856 Factor 3 4,826 1 8.70 0.005 Two-way interactions 873 3 0.52 0.668 Factors 1 and 2 169 1 0.31 0.584 Factors 1 and 3 809 1 1.46 0.235 Factors 2 and 3 11 1 0.02 0.889 Explained 13,668 6 4.105 0.003 Residual 20,531 37 Total 34,198 43 must have been an increase in the number of comparison can be made between the amino active bacteria or an increase in the activity per acids and glucose. cell to explain the faster turnover times of the The use of multivariate analysis, however, did free-floating bacteria. Inspection of the data in prove to be a fruitful approach to pursue this Table 3 reveals that the former explanation did phenomenon. The advantages in microbial ecol- perhaps occur with the amino acid mix, whereas ogy of reducing the variability of a diverse the latter explanation applied to glucose. Both of population to its most salient factors is now these mechanisms are commonplace and have becoming well documented (11, 12). The forma- been observed to operate in other aquatic envi- tion of five distinct types of aquatic environ- ronments (13, 25). Unfortunately, because it was ments based on the factors of salinity, heterotro- impossible to determine the proportions of each phic uptake, and particulate load was both amino acid taken up from the mix, no definitive fascinating and pertinent to this investigation. TABLE 7. Three-way analysis of variance of percent activity attached (glucose) by factors 1, 2, and 3 Mean Multiple Of Factor Category n score R2squaresSource of variation Sum df F Significance 1 Negative 28 34.8 0.036 Positive 16 24.2 2 Negative 17 34.8 0.012 Positive 27 28.5 3 Negative 30 20.8 0.30 Positive 14 52.7 Main effects 11,195 3 6.93 0.001 Factor 1 216 1 0.40 0.530 Factor 2 1,270 1 2.36 0.133 Factor 3 9,522 1 17.69 <0.001 Two-way interactions 1,710 3 1.06 0.378 Factors 1 and 2 1,575 1 2.93 0.096 Factors 1 and 3 291 1 0.54 0.467 Factors 2 and 3 46 1 0.09 0.770 Explained 12,905 6 4.00 0.004 Residual 19,918 37 Total 32,823 43 1236 BELL AND ALBRIGHT APPL. ENVIRON. MICROBIOL. When the percent activity of the free-floating N. and attached fraction was examined within each 00 %O cluster (Table 5), the free floaters unequivocally 8 4) 4) 0 predominated in all marine samples (clusters A z +1 +1 and B). This finding agrees well with the data of 00 numerous other workers who have found a -Q c 0 preponderance of free-floating bacteria in the sea (see above). Further investigation into the 00 o eq (7 0 0o 0o cr § effect of marine water (factor 1) on heterotrophic F - 10 O activity revealed that it operated only on the E 0 +1 +1 +1 +1 uptake of amino acids by increasing the turnover t 00 0% x I 00O Nd time per active attached cell. A probable expla- nation of this observation is in the effect of salinity on the chemiosmotic potential of the cell e-> NO membrane (15). Unfortunately the turnover time A U ~~~~+1+1 of the marine samples were not all uniformly slow as would be expected with a chemiosmotic imbalance. Cluster A comprised estuarine sites and frontal regions within the Strait of Georgia, 4) < t ..eV 't t le where turnover was comparatively fast. Why 0U) Z tl Alt co00 salinity should have had a selective effect on the 4)i -4 _ 00 activity of attached cells is unknown. The cluster where the converse situation ap- $- 0 Z ~~ +1+1 +1 +1 plied and attached bacteria constituted the domi- nant role in heterotrophic uptake was cluster D. t0 coi N 10 ON These sites all contained heavy particulate loads 4-Cu and supported the later finding that factor 3 was t- influential in effecting percent activity attached. 0 _ ~N A substrate-specific response again occurred tl^ en oo under the influence of factor 3 with glucose '4-0 eni causing an increase in the activity per cell, 4) E b whereas amino acids produced higher numbers en 0 en0 2 of active bacteria. It is interesting to speculate 4-.= that such a difference may reflect subtle changes 0 X~el o O 0 +1+1 +1 +1 v in the bacterial populations. It seems reasonable 0e 2 el; el;0 to suggest that the population adapted to amino o. acid uptake in the natural environment may be 20) engaged in proteolytic activity or the breakdown 0 of other complex molecules such as detritus. 0) 4) Such bacteria would be expected to show a close Cu 00- to association with their particulate substrate and may react to high particulate loads by increasing 00 their numbers to U U Nr N F promote colonization or form 0 aggregates. Bacteria utilizing glucose, on the 4oo o m other hand, because of the exceptionally high F- +I1+I +1 +1 on liability of the compound, may find an increase in activity a more appropriate response to the transient concentrations of glucose from previ- 4) . .C ..~ _~.> 0) W 00q ce ous reactions, most notably phytoplankton me- o O~~~no tabolism. This argument can also be extended to o~ XZ explain the 46% attached activity with amino acids compared with the 31% with glucose (Ta- ble 3). The greater the complexity of the mole- 0.4) 41) 0. C,) z cule transported the more likely the bacteria >o +l C: involved will be associated with the larger-size particulate fraction. The most notable case where substrate-specif- -1- ic responses occurred was in cluster C. This I large cluster comprised 18 sites, all of which were freshwater lakes or rivers with turnover VOL. 43, 1982 BACTERIA IN A SELECTION OF WATER BODIES 1237 times slow enough to classify them as oligotro- 8. Goulder, R. 1976. Relationships between suspended solids phic (21). Glucose again elicited a substantially and standing crops and activities of bacteria in an estuary during a Neap-Spring Neap tidal cycle. Oceologia 24:83- faster turnover time in the free-floating fraction 90. due to a shorter turnover time per free-floating 9. Goulder, R., A. S. Blanchard, P. L. Sanderson, and B. active cell. The reverse applied to the amino Wright. 1980. Relationships between heterotrophic bacte- ria and pollution in an industrialized estuary. Water Res. acids, with faster times observed in the attached 14:591-601. portion. This conforms to the pattern already 10. Hobbie, J. E., R. J. Daley, and S. Jasper. 1977. Use of described and lends credence to the above hy- Nuclepore filters for counting bacteria by fluorescent pothesis. Presumably in the oligotrophic sys- microscopy. Appl. Environ. Microbiol. 33:1225-1228. tems with the paucity of nutrients, colonization 11. Holder-Franklin, M. A. 1981. The development of biologi- cal and mathematical methods to study population shifts ofparticulates would be more critical, and hence in aquatic bacteria in response to environmental change. the different response of amino acids and glu- Scientific series no. 24. Inland Waters Directorate, Envi- cose would be more demarcated. ronment Canada, Ottawa, Canada. The of heterotrophic activity with 12. Holder-Franklin, M. A., M. Franklln, P. Cashion, C. partitioning Cormier, and L. Wuest. 1978. Population shifts in hetero- the amino acids mix and glucose has produced trophic bacteria in a tributary of the Saint John River as consistent patterns in this selection of water measured by taxometrics, p. 44-50. In M. W. Loutit and bodies. The diversity of the sites sampled should J. A. R. Miles (ed.), Microbial ecology. Springer Verlag, ensure these New York. the significance of the results for 13. Hoppe, H. G. 1976. Determination and properties of ac- two substrates. However, generalizations about tively metabolizing heterotrophic bacteria in the sea, the contribution of particle-attached bacteria investigated by means of autoradiography. Mar. Biol. versus free-floating bacteria to the heterotrophic 36:291-302. cau- 14. Jannasch, H. W. 1956. Vergleichende Bakteriologische activity of a water body have to be made untersuchung der Adsorptionswirkung des Nil-Treibsch- tiously in the light of the substrate-specific na- lammes. Ber. Limnol. Flusst. Freudenthal Munden. 7:21- ture of the uptake. Also, because the sites were 27. sampled only once between May and Novem- 15. Konings, W. H., and H. Veldkamp. 1980. Phenotypic to be responses to environmental change, p. 161-192. In D. C. ber, changes in the water with time have Ellwood, J. N. Hedger, M. J. Latham, J. M. Lynch, and ignored. Such temporal variations would obvi- J. H. Slater (ed.), Contemporary microbial ecology. Aca- ously be important in the ecology of these water demic Press, Inc., London. masses, but were not the topic of this project. 16. Paerl, H. W. 1974. Bacterial uptake of dissolved organic matter in relation to detrital aggregation in marine and Analyses have been confined to the influence of freshwater systems. Limnol. Oceanogr. 19:966-972. absolute levels of nutrients and physicochemical 17. Paerl, H. W. 1975. Microbial attachment to particles in parameters on the partitioning of heterotrophic marine and freshwater ecosystems. Microb. Ecol. 2:73- activity. 83. 18. Pomeroy, L. R. 1974. The ocean's food web, a changing ACKNOWLEDGMENTS paradigm. Bioscience 24:499-504. 19. Riemann, B. 1978. Differentiation between heterotrophic We acknowledge the financial support of the Natural Sci- and photosynthetic plankton by size fractionation, glu- ences and Engineering Research Council of Canada. cose uptake, ATP and chlorophyll content. Oikos 31:358- 367. LITERATURE CITED 20. Roper, M. M., and K. C. Marshall. 1979. Effects of salini- 1. Azam, F., and 0. Hohn-Hansen. 1973. Use of tritiated ty on sedimentation and of particulates on survival of substrates in the study of heterotrophy in seawater. Mar. bacteria in estuarine habitats. Geomicrobiol. J. 1:103-116. Biol. 23:191-196. 21. Seki, J., K. S. Shortreed, and J. G. Stockner. 1980. Turn- 2. Azam, F., and R. E. Hodson. 1977. Size distribution and over rate of dissolved organic materials in glacially- activity of marine microheterotrophs. Limnol. Oceanogr. oligotrophic and dystrophic lakes in British Columbia, 22:492-501. Canada. Arch. Hydroiol. 90:210-216. 3. Bell, C. R., and L. J. Albright. 1981. Attached and free- 22. Sieburth, J. M., R. D. Brooks, R. V. Gessner, C. D. floating bacteria in the Fraser River Estuary, British Thomas, and J. L. Tootle. 1974. Microbial colonization of Columbia, Canada. Mar. Ecol. Prog. Ser. 6:317-327. marine plant surfaces as observed by scanning electron 4. Berman, T., and M. Stiller. 1977. Simultaneous measure- microscopy, p. 418-432. In R. R. Colwell and R. Y. ment of phosphorus and carbon uptake in Lake Kinneret Morita (ed.), Effect of the ocean environment on microbi- by multiple isotopic labelling and differential filtration. al activities. University Park Press, Baltimore. Microb. Ecol. 3:279-288. 23. Strickland, J. D. H., and T. R. Parsons. 1968. A practical 5. Burnison, B. K. 1975. Microbial ATP studies. Verh. Int. handbook of sea-water analysis. Bull. Fish. Res. Bd. Can. Ver. Limnol. 19:286-290. 167:185-192. 6. Ferguson, R. L., and P. Rublee. 1976. Contribution of 24. Wiebe, W. J., and L. R. Pomeroy. 1972. Microorganisms bacteria to standing crop of coastal plankton. Limnol. and their association with aggregates and detritus in the Oceanogr. 21:141-144. sea: a microscopic study. Mem. Ist. Ital. Idrobiol. Dott 7. Geesey, G. G., and J. W. Costerton. 1979. Microbiology Marco de Marchi Pallanza Italy 29(Suppl.):325-352. of a northern river: bacterial distribution and relationship 25. Wright, R. T. 1978. Measurement and significance of to suspended sediment and organic carbon. Can. J. Micro- specific activity in the heterotrophic bacteria of natural biol. 25:1058-1062. waters. Appl. Environ. Microbiol. 36:297-305.