Dynamics of extracellular enzymatic activities in a shallow Mediterranean ecosystem (Tindari ponds, Sicily)

CSIRO PUBLISHING

www.publish.csiro.au/journals/mfr

Marine and Freshwater Research, 2005, 56, 173–188

Dynamics of extracellular enzymatic activities in a shallow Mediterranean ecosystem (Tindari ponds, Sicily) G. CarusoA,B , L. MonticelliA , F. AzzaroA , M. AzzaroA , F. DecembriniA , R. La FerlaA , M. LeonardiA and R. ZacconeA A Institute

for the Marine Coastal Environment (IAMC) – Section of Messina, National Research Council, 98122 Messina, Italy. B Corresponding author. Email: [email protected]

Abstract. Three microbial extracellular enzymes, leucine aminopeptidase (LAP), β-glucosidase (β-glu) and alkaline phosphatase (AP), were studied in six small Mediterranean littoral ponds, to evaluate the diversity of microbial activities relative to prevailing environmental conditions. The marked diversification of the trophic states, ranging from oligotrophy to eutrophy, in the ponds was reflected in a range of enzyme patterns at different spatial and temporal scales. There were higher levels and greater variability of microbial activity in the oldest and most ‘confined’ ponds (ranges: 0.55–4360.00 nm h−1 , 0.15–76.44 nm h−1 , 1.29–1600.00 nm h−1 for LAP, β-glu and AP respectively) compared with the youngest and most seaward ponds (ranges: 22.64–612.0 nm h−1 , 0.06– 48.89 nm h−1 , 0.32–744.0 nm h−1 for LAP, β-glu and AP respectively). The close relationship of the degradative potential with chlorophyll-a and particulate organic carbon could be a consequence of the stimulating effect of phytoplankton-released polymeric compounds (organic matter) and/or a response of the microbial community to warm temperatures, which were recorded from July to September. Within an area less than 1 km2 , different aquatic ecosystems coexist and maintain their distinctive properties in terms of microbial biogeochemical processes. Extra keywords: carbon cycle, microbial enzymes, phosphorus cycle, shallow waters.

Introduction The enzymatic hydrolysis of polymeric compounds by microorganisms is generally accepted as the first limiting step in the transformation of organic matter in many aquatic ecosystems.Various ecological studies have shown the importance of extracellular enzymatic activity in the utilisation and mineralisation of these substrates by planktonic communities (Cho and Azam 1988; Chrost 1991; Fuhrman 1992; Azam et al. 1994). These investigations suggest that the same activity is sensibly affected by environmental characteristics, such as the availability of inorganic nutrients, quantity and quality (in terms of lability) of organic substrates and physical and chemical variables (temperature and salinity) (Chrost 1991; Hoppe et al. 2002). The functional significance of the most frequently studied enzymes (leucine aminopeptidase, LAP, β-glucosidase, β-glu, and alkaline phosphatase, AP) is widely known in natural ecosystems: LAP and β-glu are involved in the hydrolytic decomposition of proteins and polysaccharides, respectively, and therefore in carbon cycling, whereas AP regulates phosphorus regeneration from organic phosphates, and then the phosphorus cycle (Hoppe et al. 2002). Hence, microbes are the basis of the whole living community (Hoppe 1986) and the study of bacteria–organic matter interactions is of particular interest © CSIRO 2005

in restricted areas, such as coastal lagoons, where small-scale variability of environmental parameters creates biotopes that respond differently in terms of microbial biogeochemical processes. Coastal lagoons and even small ponds on the beach are transition zones between limnetic and marine conditions. Despite the importance of brackish waters as sites of biogeochemical cycling, there have been few studies on the microbial hydrolysis of polymeric compounds through extracellular enzymatic activities (Chrost 1989; Ammerman and Azam 1991; Hoppe et al. 1998) and microbial food-web dynamics in this particular kind of environment are still poorly understood. The Tindari ecosystem is a small, shallow area of notable naturalistic value located along the Tyrrhenian coast of Sicily. This environment is a natural dynamic system characterised by strong morphological changes, with formation of littoral bars delimiting a series of coastal ponds (Leonardi and Giacobbe 2001). The ponds are not directly connected to the sea, but are separated from it by a system of littoral sand bars. The scarce exchange of waters among the ponds is probably responsible for the functional diversification of the Tindari ecosystem. Previous studies, devoted to the environmental characterisation of this ecosystem, have pointed out the various hydrobiological and trophic features 10.1071/MF04049

1323-1650/05/020173

174

Marine and Freshwater Research

of the ponds, which have been formed in different periods during the last century and are the result of a gradual isolation from the sea, in relation to the relative importance of both marine and continental water inputs (Thalassographic Reports 1990, 1991; Leonardi and Giacobbe 2001). Their very limited surface–air interface leads to a strong heterogeneity in the ponds owing to the development of microclimates. Because of their strong diversity in terms of both hydrobiological patterns and organic supply, these small ponds represent a model study area as natural mesocosms; in fact, even if covering an area less than 1 km2 , they display well-differentiated trophic conditions, with a gradient ranging from oligotrophy to eutrophy. Similar examples of high morphological diversification in the coastal region include rock pools or big lagoons (i.e. Cienaga Grande, Colombia; Hoppe et al. 1983, or Ria de Aveiro, Portugal; Hoppe et al. 1996). In the framework of the National Council Research Project ‘Cultural Heritage’, a multidisciplinary study was undertaken, aimed at safeguarding this particular area. The diverse environmental conditions existing in the Tindari ecosystems offer a good opportunity to study how a microbial community may vary its metabolic patterns at different spatial and temporal scales in response to organic inputs and physical factors. Microorganisms are known to promptly vary their enzyme profiles depending on the amount and bioavailability of organic substrates; consequently, we may expect that maximum variation in the measures of microbial decomposition activity occurs when comparing sites that differ in their trophic features. Furthermore, we would expect temporal variation in enzyme activity levels because of the positive effect of warm temperatures on the decomposition process (Patel et al. 2000). In order to verify these hypotheses, the dynamics of three microbial enzyme activities (EEA) (leucine aminopeptidase (leucyl aminopeptidase, EC 3.4.11.1, LAP), β-glucosidase (β-d-glucoside glucohydrolase, EC 3.2.1.21, β-glu) and alkaline phosphatase (EC 3.1.3.1, AP)) at mean temporal scale in relation to abiotic (physical and chemical) and biotic variables were estimated in Tindari ponds. Our study aimed to improve knowledge on carbon (C) and phosphorus (P) cycling in brackish environments.

G. Caruso et al.

N 0 100

500 m

GULF OF PATTI

C

F

D

E

B

G ITALY

A: Marinello B: Mergolo C: Verde D: Fondo porto E: Porto F: Nuovo G: Sea

Fig. 1.

A Tyrrhenian Sea

Tindari ponds with sampling stations.

matter from the surrounding lands justifies the low salinity values and supports the growth of lacustrine flora. Mergolo pond (3.8-m deep) is characterised by a negative hydrological balance and low continental inputs, which determine carbonate precipitation (Azzaro 1995). Verde pond is 3-m deep; it is influenced by continental and meteoric waters that become enriched in organic and inorganic compounds following their passage through the overhanging cliff (altitude 150 m), heavily colonised by sea-gulls (Giacobbe et al. 1990; Azzaro 1995). In this pond, particular climatic conditions (e.g. high summer temperatures) occasionally result in anoxia and fish mortality. The outermost ponds (Fondo Porto, Porto and Nuovo) are more recent and mainly influenced by seawater inflows, through infiltration mechanisms or direct contributions during storms. Fondo Porto is the shallowest pond (2.75-m deep). Sampling and parameters assayed From February 1997 to March 1998, a total of 13 surface-water samples were collected monthly in six ponds and in one marine site, which was included as a reference station (Fig. 1). Physical and chemical parameters

Materials and methods Study area The Tindari coastal ecosystem (Fig. 1), covering approximately 697 000 m2 , is a small littoral area located in the Patti Gulf (Messina, Italy), behind the Tindari Cliff (38.13◦ N 15.05◦ E), that currently includes six shallow ponds (Marinello, Mergolo, Verde, Fondo Porto, Porto and Nuovo). Their shape, number and dimensions are continuously changing owing to the rapid evolution of coastal morphology. All basins are shallow, with the maximum depth at Marinello pond (4.2 m). Marinello, Mergolo and Verde are the innermost, oldest and more typically brackish ponds. Marinello pond has no exchanges with the sea and is characterised by freshwater inflow from an underground aquifer system. The surface runoff carrying dissolved and particulate

Temperature (T◦ C) and salinity (S) measurements were carried out using an oceanographic thermometer and a salinometer respectively. Dissolved oxygen (O2 ) was estimated using the Winkler method (Carpenter 1965). Nutrient concentrations (nitrite, NO2 , nitrate, NO3 , and orthophosphate, PO4 , ions) were determined according to Genovese and Magazzù (1969); ammonia (NH4 ) concentrations were calculated according to the Aminot and Chaussepied (1983) method (detection limits: NH4 0.05–15 µm; NO2 0.01–2.5 µm; NO3 0.05–45 µm; PO4 0.03–5 µm), using a Lambda 3 spectrophotometer (Perkin-Elmer Analytical Instruments, Wallesley, MA). Trophic parameters Chlorophyll-a (chl-a), particulate organic carbon (POC) and nitrogen (PON) contents were determined as indicators of the trophic

Microbial enzymes in shallow Mediterranean ponds

conditions. The chl-a concentration was measured through extraction in 90% acetone of the material recovered on GF/F filters (Whatman International Ltd, Maidstone, Kent, UK); readings were carried out in a F-2000 spectrofluorometer (Hitachi, Osaka), previously calibrated with serial dilutions of chl-a standard from Anacistis nidulans (SigmaAldrich, St Louis, MO). The phytoplankton biomass in amounts of carbon was calculated from chl-a values, assuming a C : chl-a ratio of 40, which is the most suitable for the environment studied. For estimation of particulate organic matter (POC and PON), 500-mL samples were concentrated on precombusted Whatman GF/C filters and processed at 950◦ C in a Perkin-Elmer carbon-hydrogennitrogen (CHN)-Autoanalyzer 2400, using acetanilide as a standard (Iseki et al. 1987). Microbial parameters Samples for bacteriological analysis were collected in sterile bottles, stored at +4◦ C, and processed in the laboratory within 5 h of sampling. Microbial enzyme activities Extracellular enzyme activities (LAP, β-glu and AP) were determined according to Hoppe (1983). Briefly, triplicates of 10-mL unfiltered subsamples were incubated with a 200-µm concentration of each specific fluorogenic substrate analogue (l-leucine-4-methylcoumarinylamid hydrochloride (leu-MCA), 4-methylumbelliferyl (MUF) β-d-glucopyranoside, 4-methylumbelliferyl (MUF) phosphate (Sigma)) separately for each enzyme. From preliminary tests with increasing amounts of substrate (from 20 to 400 µm), carried out on mesotrophic and eutrophic waters, the concentration of 200 µm was determined to be the saturation substrate concentration and chosen as working concentration. The fluorescence released by substrate hydrolysis was measured using a F-2000 Hitachi spectrofluorometer as the increase between 0 time (initial time) and after 3 h incubation at the in situ temperature. Calibration curves with concentrations from 200 to 800 nm of 7-amino-4-methylcoumarin (MCA) and 4-methylumbelliferone (MUF) were performed for LAP, β-glu and AP respectively. Data were expressed in terms of potential hydrolysis rate of the substrates, as nm-leucine, nmglucoside and nm-PO4 potentially released per hour respectively. The nanomoles of leucine and glucoside potentially released by LAP and βglu were converted in nanograms of carbon using a factor of 72 and the nanomoles of phosphorus released by AP were converted to nanograms by a factor of 31. Microbial densities Bacterioplankton direct count (BDC) was estimated on 50-mL subsamples fixed in 2% formaldehyde (final concentration), filtered onto black polycarbonate filters (0.2-µm porosity; Nuclepore Inc., Pleasanton, CA), stained by DAPI (4 ,6-diamidino-2-phenylindole, Sigma) and observed using an Axioplan epifluorescence microscope (Carl Zeiss Italia SpA, Arese, Italy) as described by Porter and Feig (1980). The culturable fraction of heterotrophic aerobic marine bacteria was determined by spread-plate method on marine agar (Difco, Franklin Lakes, NJ) incubated at 20◦ C for 7 days; the fraction of culturable heterotrophic aerobic non-marine bacteria was estimated by spreading on duplicate plates of a medium without salts (peptone 5 g, yeast extract 1 g, distilled water 1000 mL, pH 7.6 ± 0.2◦ C, corrected with TRIS (hydroxymethyl)aminomethane) and incubating in the dark for 10 days at 20◦ C. For an auxiliary characterisation of the allochthonous inputs present in the area, fecal coliforms (FC) and enterococci (ENT) were determined using membrane filters according to Standard Methods (APHA 1992). Statistical analysis In order to evaluate the temporal variability of each parameter, compared with its mean value, the percentage coefficient of variation

Marine and Freshwater Research

175

(CV = (standard deviation/mean) × 100) was calculated. Analysis of variance (ANOVA) was used to investigate for statistically significant differences in the assayed parameters depending on sampling site and season. The strength of association between pairs of variables was assessed by Pearson (r) correlation coefficient. Weighted-average values of trophic variables and enzyme activities were computed using the formula: sum (difference between two following samples, in days) × (mean value of each variable obtained from two following samples).

Results Hydrological variables The annual mean values of physical and chemical variables with minimum, maximum and percentage coefficient of variation are reported in Table 1. The temporal distribution of each variable is shown in Fig. 2. Temperature ranged, on average, from 18.33 to 18.89◦ C (mean value) and followed a similar seasonal cycle in all the ponds (Fig. 2), with the minimum during January (9.78◦ C, Marinello) and the maximum during July (29.00◦ C, Nuovo). Temperature values showed a homogeneous distribution in all the ponds examined. In contrast, the six ponds were clearly differentiated from each other on the basis of salinity, which decreased along a gradient from the outermost (Nuovo, mean value = 36.21) to the innermost pond (Marinello, mean value = 19.50, the pond most influenced by freshwater input). On a temporal scale, values were widely varied, particularly in Marinello (from 3.37 to 32.00) and Verde (from 19.46 to 32.51) owing to higher contribution of continental waters; in the other ponds (Porto, Fondo Porto and Nuovo) salinity values were higher, indicating a strong influence of the sea and high homogeneity during the year (CV ∼ 10%). Evaporation occurred during summer season in all the ponds (Fig. 2). No significant differences in terms of oxygen content were found within the ecosystem; a general increase (more than 7 mL L−1 ) was measured in winter, owing to the decrease in temperature, and a decrease in summer (less than 5 mL L−1 ), except for a peak value of 9.67 mL L−1 in Verde in coincidence with a phytoplankton bloom. Trophic variables, enzyme activity patterns and microbial densities As Tindari ponds are affected by a spectrum of autochthonous and allochthonous inputs, the results of the trophic and microbiological determinations are reported by organising the different sites across a gradient depicted by their chlorophyll or organic matter contents. Therefore, in the text we will refer to different ponds grouped, according to their trophic features, into oligo-, meso- and eutrophic ponds, in order to point out the existence of distinctive patterns of microbial degradation with respect to organic polymers. Oligotrophic ponds As shown in Table 1, Mergolo and Porto ponds were representative of low degree of trophism. Mean chl-a

176

Marine and Freshwater Research

G. Caruso et al.

Table 1. Annual mean values, x, of physical, chemical and trophic parameters, with percentage coefficients of variation (CV) and ranges of variation

Temperature

(◦ C)

Salinity O2 (mL L−1 ) NH4 (µm) NO2 (µm) NO3 (µm) PO4 (µm) Chl-a (µg L−1 ) POC (µg C L−1 ) PON (µg N L−1 )

x (CV) min–max x (CV) min–max x (CV) min–max x (CV) min–max x (CV) min–max x (CV) min–max x (CV) min–max x (CV) min–max x (CV) min–max x (CV) min–max

Marinello

Mergolo

Verde

Fondo Porto

Porto

Nuovo

Sea

18.61 (34.35) 9.78–28.50 19.50 (39.13) 3.37–32.00 5.73 (18.98) 3.72–7.62 1.15 (136.52) 0.05–5.94 0.21 (78.92) 0.07–0.63 3.57 (114.98) 0.12–12.37 0.37 (67.88) 0.15–1.09 1.64 (112.75) 0.22–7.07 307.0 (62.83) 78.71–610.0 49.57 (62.37) 12.50–107.0

18.36 (35.45) 10.28–28.00 27.43 (10.33) 21.79–32.53 6.00 (12.05) 4.80–7.38 1.86 (80.64) 0.47–5.55 1.10 (114.70) 0.12–4.25 3.96 (90.59) 0.11–12.27 0.28 (37.72) 0.07–0.50 0.38 (42.66) 0.15–0.69 148.0 (56.48) 50.60–335.0 21.45 (51.46) 12.50–50.81

18.41 (35.22) 9.96–28.00 23.79 (13.80) 19.46–32.51 6.49 (22.67) 4.48–9.67 3.33 (135.53) 0.38–14.64 0.15 (39.47) 0.04–0.27 6.56 (79.61) 0.47–14.55 0.55 (102.28) 0.01–2.16 23.0 (275.99) 0.44–233.0 2230.0 (207.9) 112.0–17280.0 385.0 (211.81) 20.09–3030.0

18.52 (35.50) 9.86–28.50 34.01 (10.19) 28.00–40.08 5.84 (10.96) 4.89–7.18 0.63 (58.87) 0.23–1.50 0.13 (48.14) 0.06–0.24 4.12 (113.65) 0.36–16.25 0.27 (35.80) 0.10–0.40 1.07 (67.53) 0.17–5.78 216.0 (69.58) 90.20–548.0 30.73 (72.83) 7.25–86.00

18.33 (34.00) 10.50–28.50 33.96 (10.18) 29.39–39.26 5.93 (10.04) 4.90–7.07 0.50 (58.01) 0.09–1.04 0.17 (89.45) 0.05–0.61 1.50 (85.86) 0.21–4.49 0.24 (41.39) 0.03–0.40 0.64 (71.51) 0.15–1.46 187.0 (57.25) 76.80–419.0 26.94 (66.58) 2.00–70.63

18.89 (32.93) 11.30–29.00 36.21 (9.34) 30.26–41.00 5.66 (12.59) 4.56–7.18 0.54 (71.23) 0.01–1.40 0.06 (51.37) 0.01–0.13 2.01 (80.98) 0.25–5.64 0.24 (26.89) 0.10–0.36 1.11 (100.25) 0.18–3.32 325.0 (79.40) 100.80–920.0 51.52 (84.63) 18.10–171.0

19.92 (23.32) 14.51–28.00 37.08 (1.32) 36.46–38.04 5.72 (10.40) 4.80–6.57 0.20 (117.10) 0.01–0.80 0.06 (51.37) 0.01–0.13 1.61 (88.04) 0.21–4.97 0.24 (38.68) 0.08–0.45 0.25 (73.44) 0.15–0.85 102.0 (35.09) 66.2–170.0 14.83 (35.40) 5.60–25.00

concentrations were at their minimum level (0.38 µg L−1 , in Mergolo, followed by Porto, 0.64 µg L−1 ) and phytoplankton biomass dropped to 5.96 µg C L−1 in Mergolo in February. The oligotrophic characteristics of these ponds were also suggested by the lowest incidence of organic matter (mean POC values: 148.0 and 187.0 µg C L−1 , mean PON values: 21.45 and 26.94 µg N L−1 in Mergolo and in Porto respectively). Based on chl-a, POC and PON values, ANOVA confirmed that Mergolo was significantly different from Marinello (F1,24 = 5.966, 7.433, 9.539, P < 0.05 respectively) and Nuovo (F1,24 = 5.399, 5.516, 5.811, P < 0.05 respectively). Weighted average chl-a and POC values (Table 2) further underlined the presence of very low levels of phytoplanktonic biomass (0.015 and 0.026 g C m−3 ) and organic supplies (0.148 and 0.187 g C m−3 in Mergolo and Porto respectively). This condition of reduced trophism was not justified by the scarcity in nutrients, which were always measured in appreciable concentrations and did not drop to limitation levels, for example nitrogen was higher than 0.11 µm NO3 in Mergolo and 0.05 µm NO2 in Porto and minimum detected phosphorus was 0.07 and 0.03 µm PO4 in Mergolo and Porto respectively. Moreover, Mergolo pond was characterised by mean NO2 concentrations (1.10 µm) that exceeded by one order of magnitude those found in the other ponds (0.06– 0.21 µm, Table 1 and Fig. 2), as suggested by ANOVA values (F1,24 = 6.389, 7.338, 7.736, 6.995, P < 0.05, 8.759, P < 0.01, obtained comparing Mergolo v. Marinello, Verde, Fondo Porto, Porto and Nuovo data respectively). In Porto, the concentrations of NH4 , NO3 and PO4 were at their minimum. The low availability of organic matter of Mergolo and Porto was reflected by low LAP and β-glu values (Table 3).

In particular, β-glu activity rates in Porto reached the absolute average minimum value (0.06 nm glu h−1 in March), and differed significantly from those found in Marinello (F1,24 = 4.28, P < 0.05) and Verde (F1,24 = 5.95, P < 0.05). Coefficients of variation (LAP: 93.48 and 88.15, β-glu: 91.23 and 118.61, in Mergolo and Porto respectively) indicated quite a regular course of the degradation processes on proteins and polysaccharides in these ponds compared with the other ones. Unlike in Porto, enhanced AP rates (404.0 nm PO4 h−1 ) were measured in Mergolo and represented the most distinctive biogeochemical feature of this pond. Alkaline phosphatase rates were significantly different from those recorded in Fondo Porto, Porto and Nuovo (F1,24 = 4.55, 4.81, 4.39, P < 0.05 respectively). There was a moderate incidence of bacterioplankton in Porto and Mergolo, with similar ranges of variation (Table 3). The nutrients released by enzyme activities and the increased temperature probably supported the rise of BDC, mainly during summer months (not shown in Table 3). Within the total community, marine bacteria accounted for a low percentage (average value 1.72%) in Mergolo, reaching the absolute minimum in Porto (0.82% of BDC). Mesotrophic ponds Fondo Porto and Marinello displayed mean chl-a concentrations of 1.07 and 1.64 µg L−1 , respectively, indicating their condition as mesotrophic ponds (Table 1). Consistent organic supplies characterised these ponds (mean POC values 216.0 and 307.0 µg C L−1 in Fondo Porto and Marinello respectively). Particulate organic nitrogen distribution closely reflected that recorded for POC. Marinello was significantly

Marine and Freshwater Research

35

45

30

40

25

35

20

30 S

T (°C)

Microbial enzymes in shallow Mediterranean ponds

15

25

10

20

5

15 10

0 F M A M J J A S O N D J F M 12

F

M

A

M

J

J

A

S

O

N

D

J

F

M

F

M

A

M

J

J

A

S

O

N

D

J

F

M

16

10 NH4 (µM)

O2 (mL L⫺1)

12 8 6 4

8

4

2 0

0 F M A M J

J A S O N D J F M

16

4.2

3 2.5 NO2 (µM)

NO3 (µM)

12

8

4

2 1.5 1 0.5 0

0 F M A M J

J

A S O N D J

F M

F

M

A

M

J

J

A

S

O

N

D

J

F

M

2.5

PO4 (µM)

2.0 1.5

Marinello Fondo Porto

1.0

Mergolo Porto

Verde Nuovo

0.5 0.0 F M A M J

J

A S O N D J

Fig. 2.

F M

Monthly trend of physical and chemical variables in Tindari ponds.

Table 2. Weighted average values of trophic variables and enzyme activities found at each pond

Marinello Mergolo Verde Fondo Porto Porto Nuovo Sea

Chlorophyll-a g C m−3 year−1

Particulate organic carbon g C m−3 year−1

Leucine aminopeptidase g C m−3 year−1

β-glucosidase g C m−3 year−1

Alkaline phosphatase g P m−3 year−1

0.066 0.015 0.921 0.043 0.026 0.044 0.010

0.307 0.148 2.232 0.216 0.187 0.325 0.102

474.50 96.16 796.77 101.47 86.76 166.59 15.86

10.00 4.71 15.49 4.64 2.96 4.33 0.50

77.77 118.23 113.54 31.62 34.08 35.20 20.40

177

178

Marine and Freshwater Research

G. Caruso et al.

Table 3. Annual mean values, x, of microbiological parameters, with percentage coefficients of variation (CV) and ranges of variation

Leucine aminopep. (nm h−1 ) β-glucosidase (nm h−1 ) Alkaline phosph. (nm h−1 ) Marine bacteria (×103 CFU mL−1 ) Non-marine bacteria (×103 CFU mL−1 ) Fecal Coliforms (CFU 0.100 mL−1 ) Enterococci (CFU 0.100 mL−1 ) Bacteriopl. direct count (×108 cells L−1 )

Marinello

Mergolo

Verde

Fondo Porto

Porto

Nuovo

Sea

x (CV) min–max x (CV) min–max x (CV) min–max x (CV) min–max x (CV) min–max

739.0 (153.90) 132.77–4360.0 15.14 (116.89) 0.49–58.68 266.0 (133.96) 18.71–1089.0 22 (89) 1.1–68 1.76 (201) 0.03–9.84

145.0 (93.48) 35.00–485.0 6.94 (91.23) 0.15–17.67 404.0 (112.09) 69.03–1600.0 12 (268) 0.25–120 0.41 (301) 0.005–4.48

1210.0 (106.33) 0.55–3628.0 22.92 (116.40) 0.15–76.44 415.0 (129.40) 1.29–1548.0 77 (135) 1.2–390 0.57 (213) 0.005–3.47

159.0 (92.92) 63.47–612.0 6.94 (188.52) 0.14–48.89 110.32 (184.36) 4.84–744.0 7.6 (97) 0.36–23 0.076 (166) 0.005–0.46

134.0 (88.15) 22.64–390.0 4.44 (118.61) 0.06–18.71 117.10 (113.72) 16.13–392.0 4.2 (108) 0.45–15 0.348 (224) 0.005–2.81

257.0 (70.28) 71.94–555.0 6.53 (108.95) 0.29–19.86 121.61 (145.07) 0.32–504.0 11 (82) 0.59–30 0.28 (234) 0.005–2.28

24.86 (91.95) 6.80–76.39 0.83 (109.56) 0.00–2.99 76.45 (260.71) 6.45–696.0 3.5 (73) 0.25–7.3 0.099 (131) 0.005–0.355

x (CV) min–max x (CV) min–max x (CV) min–max

33 (225) 0–250 35 (239) 0–300 8.47 (22.85) 6.40–12.00

5 (309) 0–51 3 (256) 0–32 6.97 (34.64) 3.90–10.00

58 (124) 3–246 36 (98) 3–121 7.37 (32.83) 4.10–11.00

1.4 (188) 0–8 9 (190) 0–64 5.12 (61.75) 1.30–9.70

1.2 (222) 0–9 6 (169) 0–34 5.16 (56.49) 2.50–10.00

0.5 (168) 0–2 9 (221) 0–71 4.60 (63.25) 1.40–9.10

2 (221) 0–6 2 (129) 0–6

different from Mergolo pond on the basis of POC and PON values (F1,24 = 7.433, 9.539, P < 0.05 respectively). The detection of great amounts of POC in Nuovo (325.0 µg C L−1 ), together with high mean chl-a content (1.11 µg L−1 ), led us consider this pond as mesotrophic. The incidence of the autotrophic component and organic matter in Fondo Porto, Marinello and Nuovo was consistent with intermediate trophism, with weighted average chl-a and POC values from 0.043 to 0.066 g C m−3 and from 0.216 to 0.325 g C m−3 respectively (Table 2). Moderate amounts of nutrients, particularly abundant for NO3 (mean value, 4.12 µm) in Fondo Porto and for PO4 (mean value, 0.37 µm) in Marinello (Table 1), sustained the productive processes; conversely, in Nuovo the concentrations of inorganic nitrogen and phosphorus compounds were more similar to those detected in the oligotrophic Porto and justified the limited development of phytoplankton biomass found in this pond. As far as microbial variables are concerned (Table 3), two different trends were observed within the group of mesotrophic ponds. In Fondo Porto and Nuovo, similar mean LAP and β-glu rates and BDC values (ranging, on average, from 4.60 × 108 cells L−1 in Nuovo, to 5.12 × 108 cells L−1 in Fondo Porto) were recorded. In these outermost ponds, BDC exhibited remarkable temporal variations (CV 61.75 and 63.25 in Fondo Porto and Nuovo respectively). Alkaline phosphatase rates were of the same order of magnitude as those in Porto pond. The significant correlations detected on an annual scale between LAP, β-glu, AP v. POC, chl-a contents and BDC in both ponds (Table 4) suggested a good correspondence between organic content, autotrophic biomass, enzyme activities and bacterial abundance and therefore close relationships between production and decomposition processes.

In Marinello, LAP and β-glu (739.0 nm leu h−1 and 15.14 nm glu h−1 ) reached levels about 5 and 2 times higher than those measured in the other two ponds. The same result was observed for AP and BDC, whose rates (266.0 nm PO4 h−1 ) and values (8.47 × 108 cells L−1 ) were twice those detected in Nuovo. As observed in other innermost ponds, in Marinello, BDC values were very high (increasing to 12 × 108 cells L−1 in December) and differed significantly from those recorded in the outermost, marine, ponds (Fondo Porto, Porto and Nuovo, F1,24 = 10.65, 11.67, 15.95, P < 0.01 respectively). The percentage contribution of heterotrophic marine bacteria to BDC ranged from 1.48% in Fondo Porto, to 2.59%, in Marinello. The high continental inputs and low salinity characterising Marinello were in agreement with the highest densities of non-marine bacteria (on average, 1.76 × 103 colony-forming units (CFU) per mL, Table 3). Overall, with respect to culturable (marine + non-marine bacteria) fraction, Marinello was significantly different from Fondo Porto and Porto (F1,24 = 6.166, P < 0.05; 10.17, P < 0.01). Eutrophic ponds Verde pond was an example of eutrophic conditions. Here, chl-a content was about 20 times higher than in the other ponds and a maximum of 233.0 µg L−1 occurred in autumn (Table 1). In October, phytoplankton biomass reached a peak of 9320.0 µg C L−1 . This high phytoplankton biomass was sustained by high nutrient concentrations that reached their maxima (14.64, 14.55 and 2.16 µm for NH4 , NO3 and PO4, Table 1 and Fig. 2). Wide variations in NH4 and PO4 levels were also detected, as suggested by high CV values. The measured concentrations of POC and PON confirmed the presence in this pond of the highest organic loads, reaching a peak of 17 280 µg C L−1 and 3030 µg N L−1 in

Microbial enzymes in shallow Mediterranean ponds

Marine and Freshwater Research

179

Table 4. Pearson’s correlation coefficients calculated on annual basis for each enzyme v. environmental and biological variables, separately for each pond sampled Marinello Leucine aminopeptidase v: T S O2 PO4 NO3 NO2 NH4 Total inorganic N POC PON Chl-a Marine + non-marine bacteria β-Glu AP Total bacterioplankton β-Glucosidase v: T S O2 PO4 NO3 NO2 NH4 Total inorganic N POC PON Chl-a Marine + non-marine bacteria LAP AP Total bacterioplankton Alkaline phosphatase v: T S O2 PO4 NO3 NO2 NH4 Total inorganic N POC PON Chl-a Marine + non-marine bacteria LAP β-Glu Total bacterioplankton

Mergolo

Verde

0.75**

0.67*

Fondo Porto

Porto

0.55*

Nuovo 0.82** 0.60*

Sea

0.56*

0.60* −0.58*

0.84**

0.61* 0.61* 0.57*

0.69** 0.79** 0.93**

0.91** 0.59* 0.78*

0.87**

0.98** 0.97** 0.89**

0.85**

0.82**

−0.60*

0.91**

0.76** 0.85** 0.86** 0.81** 0.93** 0.65* 0.99**

0.64** 0.54*

0.85** 0.60*

0.85** 0.92**

0.73** 0.78** 0.80** 0.72** 0.93** 0.71** 0.94**

0.59*

0.71** −0.56* −0.75** 0.75** 0.84** 0.97**

0.84** 0.72**

0.91** 0.71** 0.82*

0.87**

0.67* 0.56*

0.59*

0.58*

0.98** 0.99** 0.74*

0.70** 0.62*

0.74**

0.62*

−0.55* −0.60*

0.71** 0.72** 0.75**

−0.62* 0.74** 0.83** 0.97**

0.72**

0.59* 0.71** 0.83*

0.97** 0.99**

0.77** 0.75** 0.74**

0.85** 0.94**

0.65* 0.71** 0.77*

Data obtained from n = 13, except for total bacterioplankton (n = 7); *P < 0.05, **P < 0.01. T, temperature; S, salinity.

October, in correspondence with a peak in chl-a content. Percentage coefficient of variation showed the high temporal variability of organic inputs. The diversity of Verde on the basis of its trophic features was also underlined by the weighted average chl-a (0.921 g C m−3 ) and POC (2.232 g C m−3 ) concentrations, largely exceeding the values calculated for the other ponds (Table 2).

The labile nature of the organic matter was suggested by the high LAP activity rates recorded in this pond (Table 3), with mean values of 1210.0 nm leu h−1 . Leucine aminopeptidase values found in Verde were also significantly different from all the other ponds (F1,24 = 8.82, 8.57, 9.02, 6.99, P < 0.01, for Mergolo, Fondo Porto, Porto and Nuovo respectively). The high CV of this enzyme showed

180

Marine and Freshwater Research

the discrete temporal variability of the proteolytic activity in this pond. Similarly to LAP, the highest mean values of AP (415.0 nm PO4 h−1 ) and β-glu (22.92 nm glu h−1 ) were detected (Table 3); these latter differed significantly from those measured in Nuovo (F1,24 = 4.52, P < 0.05) and Mergolo (F1,24 = 4.42, P < 0.05). The values of marine bacteria in Verde pond were the highest detected in Tindari ecosystem, with a peak of 3.90 × 105 CFU mL−1 in summer (Table 3). According to its culturable (marine + non-marine bacteria) fraction, Verde differed significantly from Fondo Porto, Porto and Nuovo (F1,24 = 5.755, 6.341, 5.106, P < 0.05). Generally, BDC values reflected the trophic gradients of the studied ponds; nevertheless, in Verde pond, total abundances were rather low and seemed to contrast with what was expected (i.e. a direct relationship between phytoplankton primary production and bacterial production in terms of new biomass, as reported by Cole et al. 1988). The values were similar to those typical for nutrient-poor, oceanic waters, despite the highest chl-a content; this could derive from the fact that autotrophic populations (see the high chl-a values, about 23 µg L−1 ) exceed heterotrophic ones. It is well known that bacteria dominate total microbial biomass in oligotrophic waters, whereas in eutrophic waters, where larger-size phytoplankton dominate, the bacterial dominance decays (Duchlow and Carlson 1992). Anyhow, in this study case, the bacterial abundances were consistent with previous findings from other coastal and estuarine areas with high nutrient load (Turk et al. 1992; La Ferla et al. 2002). Puddu et al. (1998) hypothesised a bacterial loss owing to bacteriovory or virioplankton lysis. Reference station At the marine station, the minimum LAP value (24.86 nm leu h−1 , on average) was measured (Table 3), and annual mean LAP data differed significantly from those measured in all the ponds (F1,24 = 5.12, 9.91, 11.02, 10.45, 10.67, 21.13, P < 0.01, compared with Marinello, Mergolo, Verde, Fondo Porto, Porto and Nuovo respectively). β-Glucosidase activity was, on average (0.83 nm glu h−1 ), significantly lower than that measured in all the ponds (F1,24 = 8.51, 12.03, 8.92, 6.09, 8.34, P < 0.01, compared with Marinello, Mergolo, Verde, Porto and Nuovo respectively), whereas relatively high AP levels (mean value = 76.45 nm PO4 h−1 ) were detected, which were significantly different from those measured in Mergolo and Verde (F1,24 = 5.69, 4.54, P < 0.05 respectively). Temporal behaviour of enzyme activities and microbial abundances The monthly patterns of the enzyme activities are shown in Fig. 3 and Fig. 4. Leucine aminopeptidase distribution (Fig. 3) revealed the presence of two separate peaks, occurring mostly in June (4360.0 and 3628.0 nm leu h−1 in Marinello and in Verde respectively) and from August to October. The spring

G. Caruso et al.

peak was, in general, higher than the autumn peak. The monthly course of β-glu was similar to that of LAP, increasing everywhere during the summer months (June–September) (Fig. 3). In this period, LAP and β-glu dynamics appeared closely associated with chl-a (r = 0.81 and 0.74; P < 0.01, n = 24, overall data respectively) and POC contents (r = 0.84 and 0.82, P < 0.01, n = 24 respectively). Peaks in the temporal distribution of AP (Fig. 4) were observed mainly in June (1457.0 nm PO4 h−1 in Verde) and in August–September, except for Mergolo and Nuovo ponds, where they occurred earlier (in May, 1600.0 nm PO4 h−1 in Mergolo, absolute maximum value). In some ponds, such as Verde in August, and Marinello, Porto and Fondo Porto in July, an abrupt fall in AP activity was observed concurrently with the peaks of bio-available PO4 . Culturable bacteria, mainly composed of heterotrophic marine bacteria (Table 3), pointed out a quite homogeneous abundance in the ponds, with slight seasonal increases in the summer–autumn period (not shown in Table 3). The lowest counts of marine bacteria were recorded in November–December (minimum 2.50 × 102 CFU mL−1 in Mergolo). High values of non-marine bacteria (ranging from 5 × 100 CFU mL−1 , in Mergolo, to 9.8 × 103 CFU mL−1 in Marinello pond) occurred during autumn–winter. Faecal bacteria (faecal coliforms and enterococci), which were always found in low concentrations in the ponds as well as at the sea, followed a similar seasonal distribution. In Marinello, inverse correlations were observed between salinity and FC, considered as bacterial indicators of faecal contamination, as well as between salinity and ENT (r = −0.68 and −0.69, n = 13, P < 0.01 respectively). The positive correlations between non-marine bacteria and POC (r = 0.62, n = 13, P < 0.05) and PON (r = 0.66, n = 13, P < 0.05) and the correspondence of peak values of FC, ENT and non-marine bacteria (data not reported) clearly indicated that the input of exogenous particle materials from the drainage of rural surroundings was responsible for the occurrence of faecal bacteria during the rainy season (October–January). An evident input of exogenous particulate matter, faecal and non-marine bacteria were also observed in Verde pond, suggesting the greater importance of the impact of the wild fauna (i.e. sea-gulls nests) and continental runoff on bacterial loading in these two ponds. In oligo- or mesotrophic environments, such as Mergolo, Fondo Porto, Porto and Nuovo, total BDC were well correlated on an annual scale with LAP, β-glu and AP, whereas significant correlations were always lacking in Marinello and Verde, nutrient-rich environments (Table 4). For data grouped on a seasonal scale, significant correlations linked BDC to LAP and β-glu (r = 0.57, 0.64, n = 24, P < 0.01 respectively), suggesting that in winter, bacterial growth depended on microbial degradation, whereas in spring it was more significantly linked to POC and chl-a (r = 0.82, 0.81, n = 18 respectively). In Marinello, Porto and Nuovo, seasonally averaged BDC values were significantly related to POC and POC

400

1000

200

0

0 F M A M J J A S O N D J F M

5000

Verde

17288

LAP (nM h⫺1)

4000

4000 3000

3000 2000 2000 1000

1000

0

0 F M A M J J A S O N D J F M

LAP (nM h⫺1)

5000

Porto

1000

4000

800

3000

600

2000

400

1000

200

0

0 F M A M J J A S O N D J F M

Sea

1000

4000

800

3000

600

2000

400

1000

200 0

0

3000

600

2000

400

1000

200 0

0 F MAM J J A S ON D J F M

5000

Fondo Porto

1000

4000

800

3000

600

2000

400

1000

200 0

0 FMAM J J A S OND J FM

5000

Nuovo

1000

4000

800

3000

600

2000

400

1000

200

0

0 F MAM J J A S ON D J F M

F M A M J J A S O N D J F M

␤-Glu (nM h⫺1); POC (µg C L⫺1)

LAP (nM h⫺1)

5000

800

␤-Glu (nM h⫺1); POC (µg C L⫺1)

2000

4000

␤-Glu (nM h⫺1); POC (µg C L⫺1)

600

1000

Mergolo

181

␤-Glu (nM h⫺1); POC (µg C L⫺1)

3000

LAP (nM h⫺1)

800

LAP (nM h⫺1)

4000

5000

LAP (nM h⫺1)

1000

␤-Glu (nM h⫺1); POC (µg C L⫺1)

Marinello

␤-Glu (nM h⫺1); POC (µg C L⫺1)

LAP (nM h⫺1)

5000

Marine and Freshwater Research

␤-Glu (nM h⫺1); POC (µg C L⫺1)

Microbial enzymes in shallow Mediterranean ponds

LAP ␤-Glu POC

Fig. 3. Temporal distribution of particulate organic matter (POC), leucine aminopeptidase (LAP), β-glucosidase (β-glu), in Tindari ponds. Enzyme values are reported in nm – leucine or glucoside relased per hour.

with LAP, confirming that the proteolytic activity over POC sustained, in turn, the bacterial growth. Activity–biomass relationship In order to correlate enzyme activity rates with the metabolically active fraction of the bacterial community, which plays a role in biogeochemical processes, the viable bacterial fraction (VBF) was determined as the percentage of heterotrophic (marine + non-marine) bacteria over total BDC. Approaches

for assessing bacterial viability are usually based on the demonstration of culturability or metabolic activity. Viable bacterial fraction represented, on average, a percentage ranging from 0.6 (in Mergolo) to 5.4 (in Verde), discarding an extremely high value (about 60%) recorded in Verde pond. This could be surprising, taking into account that in seawater, living bacteria, as determined by estimates of culturable cells, commonly account for only 0.001–0.2% of total direct counts. However, recent studies in the marine pelagic environment by

Marine and Freshwater Research

2.5

2.0

1000

2.0

1000

1.5

100

1.5

100

1.0

10

1.0

10

0.5

1

0.5

1

0.0

0

2.5

1000

2.0

1000

1.5

100

1.5

100

1.0

10

1.0

10

0.5

1

0.5

1

0.0

0

10000

Fondo Porto

0

0.0

F M A M J J A S O N D J F M

F M A M J J A S O N D J F M

Nuovo

1000

2.0

1000

1.5

100

1.5

100

1.0

10

1.0

10

0.5

1

0.5

1

0

0.0

0.0 F M A M J J A S O N D J F M

10000

2.0

1000

1.5

100

1.0

10

0.5

1

0.0

AP Chl-a (µg L

Sea

0 F M A M J J A S O N D J F M

(nM h⫺1); ⫺1)

2.5

PO4 (µM)

2.0

AP (nM h⫺1); Chl-a (µg L⫺1)

2.5

Porto

AP (nM h⫺1); Chl-a (µg L⫺1)

10000

10000

2.5

AP (nM h⫺1); Chl-a (µg L⫺1)

PO4 (µM)

10000

Verde

AP (nM h⫺1); Chl-a (µg L⫺1)

PO4 (µM)

0 F M A M J J A S O N D J F M

2.0

2.5

PO4 (µM)

10000

0.0

F M A M J J A S O N D J F M

PO4 (µM)

Mergolo

AP (nM h⫺1); Chl-a (µg L⫺1)

Marinello

PO4 (µM)

10000

2.5

PO4 (µM)

G. Caruso et al.

AP (nM h⫺1); Chl-a (µg L⫺1)

182

PO4

AP

Chl-a

0 F M A M J J A S O N D J F M

Fig. 4. Annual course of orthophosphate (PO4 ), alkaline phosphatase (AP), and chlorophyll-a (chl-a), in the six ponds examined. Enzyme values are expressed in nm of PO4 released per hour.

Bianchi and Giuliano (1996) have shown that viable bacteria could contribute to an important (frequently exceeding 20%) percentage of the total bacterial population, particularly in superficial water masses. In coastal eutrophic waters of the northernAdriatic Sea, culturable bacteria accounted for 0.1–22% of BDC (Zaccone et al. 2002). Moreover, we have to consider that the hydrological properties of our ecosystem were more similar to lacustrine rather than to marine environments. Lacustrine environments, in fact, frequently show a fraction of active bacteria larger than marine ecosystems at comparable chl-a and DOC concentrations (Haglund

2004), even though not all ‘active’ cells are actually ‘culturable’ cells. In addition, we think that the VBF found in Verde was presumably related to the highest availability in this pond of dissolved organic substrates (i.e. algal exudates) in a form prone to microbial degradation; this hypothesis should be supported by studies (Del Giorgio and Scarborough 1995) that report the fraction of metabolically active bacteria to increase with lake productivity. Sommaruga and Conde (1997) determined the presence of active cells in proportions ranging from 17% in winter to 100% in autumn in a hypertrophic lake.

Microbial enzymes in shallow Mediterranean ponds

600

7

500

6

2000

3

0

5000

M

J

J

A

D

2

1

100

1

0

0

0 F

Verde

70

700

60

600

50

500

3000

40

2000

30

1000

M

J

J

A

D

J 20

Fondo Porto

400 10 300

20

200

10

100 0

0 M

J

J

A

D

J

0 F

8 7

500

6 5

400

4 300

3

700

nM h⫺1

Porto

600

VBF

F

nM h⫺1

3

200

J

4000

700

4

300

2

nM h⫺1

F

5

400

M

J

J

A

D

J

8

Nuovo

600

7

500

6 5

400

4

300

3

200

2

200

2

100

1

100

1

0

0

0

F

M

J

J

A

D

J

VBF

4

8

Mergolo

VBF

5

3000

nM h⫺1

6

1000

nM h⫺1

700

7

VBF

nM h⫺1

4000

8

183

VBF

Marinello

VBF

5000

Marine and Freshwater Research

0 F

M

J

J

A

D

J

LAP ⫹ ␤-Glu VBF

Fig. 5. Comparison between leucine aminopeptidase (LAP) plus β-glucosidase (β-glu) activity values and the viable bacterial fraction (VBF), estimated as the percentage of heterotrophic (marine + non-marine) bacteria over bacterioplankton direct count (BDC).

The comparison between LAP and β-glu activities and VBF (Fig. 5) highlighted the occurrence of the peak of metabolic activity earlier than that of bacterial growth, mainly in June and August, suggesting that the release of simple molecules from polymers was temporally uncoupled with the increase in VBF. Discussion Coastal aquatic ecosystems are among the most geochemically and biologically active areas of the biosphere and play a considerable role in the global biogeochemical cycle. In microbial ecology, many studies shown the importance of microbial processes in biogeochemical cycles (Azam et al. 1994; Chrost 1991; Hoppe et al. 2002), but only a few studies (Chrost 1989; Kisand et al. 2001) have focused

on the relative importance of microbial activity within the ecological equilibrium of shallow coastal areas. Our investigation represents a contribution to knowledge about this topic, elucidating how and to what extent the dynamics of some enzyme activities involved in carbon and phosphorus cycles are affected by environmental factors in a small Mediterranean area. In fact, as a consequence of the intrinsic conditions (organic loads, nutrient availability) of each pond, variations in the patterns of both production and decomposition processes occur. This feature is particularly true in restricted basins, where more intense microbial activities and wide nutrient variations closely correlate with the photosynthetic production. In Tindari ecosystem, the physiological response of the microbial community to the organic matter available varied significantly, as suggested by ANOVA values, in relation to sampling site and season. Enzyme levels

184

Marine and Freshwater Research

were consistent with the trophic state of the system; in fact, high microbial activity and great temporal variability characterised the oldest, typically brackish, ponds (Marinello and Verde), where enzyme production was strongly induced by the trophic inputs. In contrast, in the youngest ponds, which were subjected to marine influence and low organic loads, low levels of enzyme activities and reduced temporal variability were detected. The marine station always showed low enzyme concentrations, with CV similar to, or slightly lower than, in the ponds. The high levels of microbial enzyme activities measured at meso- and eutrophic ponds showed that the hydrolytic potential for organic matter processing responded positively to the trophic gradient, as observed by Karner et al. (1992) and Hoppe et al. (1998). In summer, the percentage of culturable bacteria and the rates of decomposition were enhanced, suggesting changes in the physiological state of the bacterial population. Our enzyme activity values were higher than data (ranging from 12.92 to 311.11 nm h−1 for LAP) measured in the brackish Kiel Fjord (Hoppe 1986), but similar to those observed by Hoppe et al. (1998) in a eutrophication gradient in Schlei Fjord (ranging from 176 to 3041 nm h−1 and from 32.2 to 76.4 nm h−1 for LAP and β-glu respectively). They were also consistently higher than those reported in marine areas (Caruso and Zaccone 2000; La Ferla et al. 2001). Our estimates of microbial enzyme activities confirmed that within small environments, the biological processes were probably modulated by a variety of local environmental parameters. Temperature positively affected the behaviour of metabolic processes (Table 4), and a seasonal signal in enzyme patterns, with enhanced activity rates during spring– summer, was evident in all the ponds. Other studies have pointed out the importance of temperature in controlling enzyme expression in coastal ecosystems (Patel et al. 2000). Nevertheless, in our case, although correlation analysis does not allow inferences concerning cause–effect relationships, results suggested that, other than temperature, trophic supplies, in terms of chl-a and POC amounts, played a major role in modulating enzyme dynamics, in agreement with Taylor et al. (2003). Linear regression analysis performed on data pooled on a seasonal scale (data not shown) showed that chl-a content accounted for 60–76% of LAP variability in spring and autumn, respectively, and for 54–55% of β-glu variability in spring and summer respectively. In autumn, variations in enzyme levels were also ascribable to those of POC content (79 and 41% for LAP and β-glu). Variations in bacterial abundance could also justify variations detected in spring and in summer, both in LAP (45 and 32%) and β-glu levels (28 and 45%). The influence of the trophic characteristics in driving biogeochemical cycles, however, was evident in some ponds (Verde, Fondo Porto and Nuovo, Table 4) and/or seasons only. In other words, microbial activities responded not only to the eutrophication gradient, but also to external forcing

G. Caruso et al.

(i.e. storm events, occasional continental inputs), that may become important concurrently with some circumstances only and account for the pronounced variability observed in microbial degradation rates. In Verde pond, the highest enzyme levels were recorded in coincidence with the organic substrate supply originating from autotrophic production, particularly evident in this site. The dependence of bacterioplankton on the trophic sources was also stated by Kisand et al. (2001) in a large, shallow eutrophic lake. The availability of organic substrates probably stimulated the extracellular enzyme production, as already hypothesised by Nausch et al. (1998). The relationships recorded in the ponds between LAP or β-glu and chl-a and POC during summer reflected the enhanced microbial attack and cleavage of polymeric substrates, mainly proteins and polysaccharides, released by the phytoplankton breakdown (Chrost et al. 1989). This confirms that estimates of enzyme activities may provide a measure of microbial response to the supply of organic substrates bioavailable to the microbial community, in agreement with Chrost and Rai (1993). The high CV values calculated for enzymes also suggested that metabolic patterns of microorganisms reflect variations in the composition and availability of organic polymers. Of course, our monthly sampling scale may not be adequate to evaluate changes in bacterial physiological responses, usually occurring at shorttime scales (i.e. less than one day, see Andersen-Elvehøy and Thingstad 1991; Karner and Rassoulzadegan 1995). Nevertheless, enzyme profiles may provide useful insights on functional changes in aquatic environments (Hoppe et al. 2002). As observed in other coastal ecosystems (Karner and Rassoulzadegan 1995; Puddu et al. 2000), in Tindari ponds, the peak of hydrolytic activity on freshly produced organic matter was temporally shifted compared with the increase in the culturable fraction, indicating a delay between the induction of enzymes and their synthesis and release. Carbon cycle Through extracellular enzyme activities, allochthonous and autochthonous dissolved and particulate compounds are converted to simpler monomers, becoming available to bacteria. In agreement with other Mediterranean areas (La Ferla et al. 2004), in Tindari ponds, the detrital fraction largely contributed to the POC pool (representing, on average, 77.75% of POC, data not reported), and offered a potential substrate for bacterial colonisation; detritus probably originated from bottom resuspension, owing to the shallowness of the ponds. Both the two microbial enzyme activities (LAP and β-glu) examined in our study play a key role in carbon regeneration from the organic carbon pool, supporting microbial growth. The weighted average values of LAP and β-glu (Table 2) underlined the functional heterogeneity of

Microbial enzymes in shallow Mediterranean ponds

the Tindari ecosystem, where high (Marinello and Verde), medium (Mergolo, Fondo Porto, Porto and Nuovo) and low (marine station) activity levels co-exist. In Verde pond, where all enzyme activities reached their maximum expression, the amount of carbon potentially liberated as leucine by LAP was 796.77 g C m−3 year−1 . Leucine aminopeptidase always exceeded β-glu, suggesting a preferential attack by microbes on proteins, which are more labile, rather than on carbohydrates. The LAP : βglu ratio oscillated between 3.18, in February in Verde, and 876.25 in March in Porto; on seasonally averaged values, a lower ratio was observed in summer in association with the highest chl-a and POC, two organic substrates rich in polysaccharides. From our estimates of hydrolytic rates, the potential amount of carbon flowing through the microbial community was estimated; the carbon fraction potentially mobilised by microbial activities may provide an index of the available source to the planktonic community. Assuming that the bulk of the organic matter was globally degradable and bioavailable, from data reported in Tables 1 and 3, we computed that the coupled activities of LAP and β-glu were able to potentially mobilise, per day, a fraction of 217.44 and 229.04% of POC (in Verde and in Mergolo respectively), reaching a maximum in Marinello pond (469.65% per day), because of the greater incidence of both LAP and β-glu. Two possible explanations for these high percentages that largely exceeded the amount of POC present are the following: (i) the dissolved fraction of the organic matter, unfortunately not analysed in this study but usually 10 times higher than particulate matter in aquatic environments (Carlson et al. 1998), may play a role greater than particulate fraction, supplying an important source of carbon to bacterial needs; and (ii) our estimates referred to ‘potential’ decomposition rates, determined under substrate excess (experimental conditions), which may significantly differ from in situ decomposition rates. Daily rates of organic carbon hydrolysis were greatly affected by the amount of bioavailable polymers; they largely varied, being faster in Marinello (0.47 days on average, data not shown), with a minimum in June, and slower in Porto. In Porto pond, the low enzyme activity levels, and therefore a low degradation potential, resulted in a long POC potential hydrolysis time (1.33 days on average), in particular in July, similar to the sea. Phosphorus cycle The synthesis of AP is known to play an important role in the regeneration of inorganic phosphorus through the breakdown of organic phosphorus esters. By this metabolic pathway, over 99% of phosphorus required by primary producers is recycled, satisfying phosphorus demand by phytoplankton and bacterioplankton (Chrost and Overbeck 1987; Chrost 1991). The weighted average values of AP (Table 2) distinguished

Marine and Freshwater Research

185

the ponds into two groups with different metabolic patterns: (i) one exhibiting high enzyme activity (Marinello, Mergolo and Verde, where the amount of inorganic phosphorus potentially liberated from organic phosphates by AP was 113.54 g P m−3 year−1 ); and (ii) the other (Fondo Porto, Porto and Nuovo), with AP levels about one third lower. In the latter, the availability of nutrients in a mineralised form (i.e. inorganic phosphorus) may be itself sufficient to satisfy the phosphorus demand of the plankton community and could be responsible for the lack of stimulation of AP synthesis. On the contrary, the high levels of AP activity detected in Mergolo pond suggested a faster remineralisation of organic phosphates than in other ponds (Marinello and Verde), which was not reflected in a conspicuous phytoplankton biomass, as shown by chl-a values (Table 1). This suggests that factors other than the availability of PO4 ions probably controlled the autotrophic growth in this pond. It is likely that the high carbonate precipitation, characterising Mergolo pond, acted as an important mechanism of phosphorus burial in the sediments, with important implications in phosphorus dynamics (Gonsiorczyk et al. 1998). In these conditions, we may suppose that the binding of phosphorus to the sediment prevented its biological utilisation, and consequently strongly induced the synthesis of AP. In a eutrophic pond, Chrost et al. (1989) pointed out that both phytoplankton and bacterioplankton were actively involved in the synthesis of AP during phytoplankton blooms. A positive significant relationship (r = 0.83, n = 7, P < 0.05) was observed between AP and BDC in Mergolo (Table 4); this finding, together with the lack of correlation between AP and chl-a, was indicative of the bacterial origin of AP in this pond. Conversely, in Fondo Porto and Nuovo, the significant correlations found between AP and chl-a (r = 0.97 and 0.74, n = 13, P < 0.01 respectively), suggested that this enzyme activity was mainly associated with autotrophic cells and led us to suppose that phytoplankton also significantly contributed to phosphorus release for its metabolism. The relationship between AP and chl-a could also indicate the availability of phytoplankton-derived dissolved compounds, providing a suitable substratum that, in turn, stimulates the enzyme synthesis. Since the estimation of AP activity may be used to assess the phosphorus deficiency of the microbial assemblage (Jones 1997), during spring, AP activity rates seemed to be indicative of phosphorus deficiency owing to the enhanced consumption by phytoplankton. Conversely, in autumn– winter period, the levels of inorganic phosphate were higher, suggesting the change towards a phosphorus sufficiency condition. This could probably depend on the release of phosphorus in these shallow waters, through continental erosion, runoff inputs or re-suspension from sediments caused by strong winds or storms. As previously stated for mesotrophic lakes by Siuda and Chrost (1999), in the Tindari ponds, no direct relationship

186

Marine and Freshwater Research

linking AP to PO4 concentrations was observed, suggesting that, in this particular ecosystem, the rate of enzymatic decomposition of organic phosphorus compounds does not always depend upon the degree of phosphorus limitation of the environment. In other words, we agree with Chrost’s (1991) finding that the synthesis of AP was probably regulated by the microbial internal phosphorus pool, which may not reflect the phosphorus pool outside microbial cells. Moreover the lack of a correlation between organic phosphate substrate and PO4 led us to support the importance of AP in providing not only a phosphorus source, but also a carbon source (Hoppe et al. 2002). Owing to their high activity and density, bacteria play a critical role in the decomposition of organic matter and nutrient cycling in the Tindari ecosystem. The response of the microbial community to organic loading is dynamic; our estimates of the microbial extracellular activities confirmed the complexity of the microbial processes in heterogeneous environments, such as coastal ponds, where hydrological and trophic conditions are continuously changing (Zaccone et al. 2000). As it is known, microbial settlement is the result of the interaction of both abiotic and biotic variables (Chrost 1991; Zaccone and Caruso 2002) and microbes rapidly respond to variation in environmental parameters. Physical, biological and chemical gradients induce changes in the metabolic profiles of the bacterial communities, which, in turn, affect the extent of organic matter processing, as observed even in estuarine regions (Schultz and Ducklow 2000). Bacteria– phytoplankton relationships also add another source of variability of biological processes, especially in eutrophic systems. All these considerations may explain difficulties that have arisen in the understanding of microbial processes occurring in particular coastal environments. In conclusion, a strong spatial and temporal heterogeneity characterised all the abiotic and biotic parameters studied in the area: first, a result of the distance of each pond from the sea and second, to the diversity of organic inputs. The micro-environmental variability is expressed in a great variety of consistently different biological conditions that coexist in the Tindari coastal system and maintain their distinctive properties in spite of their spatial proximity. Acknowledgments The authors wish to thank Professor H.-G. Hoppe (Institut fur Meereskunde, Kiel, Germany) for the critical revision of this paper and for stimulating suggestions that greatly improved this manuscript. References APHA (1992). ‘Standard Methods for the Examination of Water and Wastewater.’ 18th edn. (American Public Health Association, Inc: Washington, DC.)

G. Caruso et al.

Aminot, A., and Chaussepied, M. (1983). ‘Manuel d’analyses chimiques en milieu marin.’ (Centre National pour l’Exploration des Oceans: Brest, France.) Ammerman, J. W., and Azam, F. (1991). Bacterial 5 -Nucleotidase activity in estuarine and coastal marine waters – Characterization of enzyme activity. Limnology and Oceanography 36, 1427–1436. Andersen Elvehøy, I. L., and Thingstad, T. F. (1991). Detection of limiting factors for bacterial activity using protease activity. Distribution and activity of microorganisms in the sea. Kieler Meeresforschungen Sonderheft 8, 392–398. Azam, F., Smith, D. C., Steward, G. F., and Hagstrom, A. (1994). Bacteria-organic matter coupling and its significance for oceanic carbon cycling. Microbial Ecology 28, 167–179. doi:10.1007/ BF00166806 Azzaro, F. (1995). Osservazioni biennali sull’ecosistema degli stagni costieri di Oliveri-Tindari (Messina): nutrienti e clorofilla a. In ‘Atti del VI Congresso della Società Italiana di Ecologia, Venezia’. (Eds O. Ravera and A. Anelli.) pp. 495–497. (Società Italiana di Ecologia: Parma, Italy.) Bianchi, A., and Giuliano, L. (1996). Enumeration of viable bacteria in the marine pelagic environment. Applied and Environmental Microbiology 62, 174–177. Carlson, C. A., Ducklow, H. W., Hansell, D. A., and Smith, W. O., Jr (1998). Organic carbon partioning during spring phytoplankton blooms in the Ross Sea polynya and the Sargasso Sea. Limnology and Oceanography 43, 375–386. Carpenter, J. H. (1965). The accuracy of the Winkler method for the dissolved oxygen analysis. Limnology and Oceanography 10, 135–140. Caruso, G., and Zaccone, R. (2000). Estimates of leucine aminopeptidase activity in different marine and brackish environments. Journal of Applied Microbiology 89, 951–959. doi:10.1046/ J.1365-2672.2000.01198.X Cho, B. C., and Azam, F. (1988). Major role of bacteria in biogeochemical fluxes in the ocean’s interior. Nature 332, 441–443. doi:10.1038/332441A0 Chrost, R. J. (1989). Characterization and significance of β-glucosidase activity in lake water. Limnology and Oceanography 34, 660–672. Chrost, R. J. (1991). Environmental control of the synthesis and activity of aquatic microbial ectoenzymes. In ‘Microbial Enzymes inAquatic Environments’. (Ed. R. J. Chrost.) pp. 29–59. (Springer-Verlag: New York.) Chrost, R. J., and Overbeck, J. (1987). Kinetics of alkaline phosphatase activity and phosphorus availability for phytoplankton and bacterioplankton in lake Plubsee (North German eutrophic lake). Microbial Ecology 13, 229–248. Chrost, R. J., and Rai, H. (1993). Ectoenzyme activity and bacterial secondary production in nutrient-impoverished and nutrient-enriched freshwater mesocosms. Microbial Ecology 25, 131–150. Chrost, R. J., Munster, U., Rai, H., Albrecht, D., Witzel, P. K., and Overbeck, J. (1989). Photosynthetic production and exoenzymatic degradation of organic matter in the euphotic zone of a eutrophic lake. Journal of Plankton Research 11, 223–242. Cole, J. J., Findlay, S., and Pace, M. L. (1988). Bacterial production in fresh and saltwater ecosystems: a cross-system overview. Marine Ecology Progress Series 43, 1–10. Del Giorgio, P. A., and Scarborough, G. (1995). Increase in the proportion of metabolically active bacteria along gradients of enrichment in freshwater and marine plankton: implications for estimates of bacterial growth and production rates. Journal of Plankton Research 17, 1905–1924. Duchlow, H. W., and Carlson, C. A. (1992). Oceanic bacterial production. In ‘Advances in Microbial Ecology. Vol. 12’. (Ed. K. C. Marshall.) pp. 113–181. (Plenum Press: New York.)

Microbial enzymes in shallow Mediterranean ponds

Fuhrman, J. A. (1992). Bacterioplankton roles in cycling of organic matter: the microbial food web. In ‘Primary Productivity and Biogeochemical Cycles in the Sea’. (Eds P. G. Falkowski and A. D. Woodhead.) pp. 361–383. (Plenum Press: New York.) Genovese, S., and Magazzú, G. (1969). ‘Manuale d’analisi per le acque salmastre.’ (Editrice Universitaria: Messina.) Giacobbe, S., Leonardi, M., Azzaro, F., and Rinelli, P. (1990). Descrizione di un esempio di “confinamento”: l’area lagunare di Oliveri-Tindari (Messina). Oebalia Supplemento XVI-II, 675–678. Gonsiorczyk,T., Casper, P., and Koschel, R. (1998). Phosphorus-binding forms in the sediment of an oligotrophic and an eutrophic hardwater lake of the Baltic lake district (Germany). Water Science and Technology 37, 51–58. doi:10.1016/S0273-1223(98)00055-9 Haglund,A.-L. (2004).Attached bacterial communities in lakes- habitatspecific differences. In ‘Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1029’. pp. 1–35. (Acta Universitatis Upsaliensis: Uppsala.) Hoppe, H. G. (1983). Significance of exoenzymatic activities in the ecology of brackish water: measurements by means of methylumbelliferyl-substrates. Marine Ecology Progress Series 11, 299–308. Hoppe, H. G. (1986). Relations between bacterial extracellular enzyme activities and heterotrophic substrate uptake in a brackish water environment. In ‘Proceedings of the 2ème GERBAM – CNRS Colloque International de Bactériologie marine, Brest’. pp. 119–128. (IFREMER: Plouzané, France.) Hoppe, H. G., Gocke, K., Zamorano, D., and Zimmermann, R. (1983). Degradation of macromolecular organic compounds in a tropical lagoon (Cienaga Grande, Colombia) and its ecological significance. Internationale Revue Der Gesamten Hydrobiologie 68, 811–824. Hoppe, H. G., Gocke, K., and Alcantara, F. (1996). Shifts between autotrophic and heterotrophic processes in a tidal lagoon (Ria de Aveiro, Portugal). In ‘Special Issue: Advances In Limnology’. Archiv Für Hydrobiologie 48, 39–52. Hoppe, H. G., Giesenhagen, H. C., and Gocke, K. (1998). Changing patterns of bacterial substrate decomposition in a eutrophication gradient. Aquatic Microbial Ecology 15, 1–13. Hoppe, H. G., Arnosti, C., and Herndl, G. F. (2002). Ecological significance of bacterial enzymes in the marine environment. In ‘Enzymes in the Environment: Activity, Ecology, and Applications’. (Eds R. G. Burns and R. P. Dick.) pp. 73–107. (Marcel Dekker: Basel, Switzerland.) Iseki, K., MacDonald, R. W., and Carmack, E. (1987). Distribution of particulate matter in the south-eastern Beaufort Sea in late summer. In ‘Proceedings NIPR Symposium Polar Biology’. pp. 35–46. Jones, R. D. (1997). Phosphorus cycling. In ‘Manual of Environmental Microbiology’. (Ed. C. J. Hurst.) pp. 343–348. (ASM Press: Washington, DC.) Karner, M., and Rassoulzadegan, F. (1995). Extracellular enzyme activity: indicators for high short-term variability in a coastal marine ecosystem. Microbial Ecology 30, 143–156. Karner, M., Fuks, D., and Herndl, G. (1992). Bacterial activity along a trophic gradient. Microbial Ecology 24, 243–257. doi:10.1007/BF00167784 Kisand, V., Tuvikene, L., and Noges, T. (2001). Role of phosphorus and nitrogen for bacteria and phytoplankton development in a large shallow lake. Hydrobiologia 457, 187–197. doi:10.1023/A:1012291820177 La Ferla, R., Zaccone, R., Caruso, G., and Azzaro, M. (2001). Enzymatic activities and carbon flux through the microbial compartment in the Adriatic Sea. In ‘Mediterranean Ecosystems: Structures and Processes’. (Eds F. M. Faranda, L. Guglielmo and G. Spezie.) pp. 485–493. (Springer-Verlag Italia: Milano.)

Marine and Freshwater Research

187

La Ferla, R., Azzaro, M., Zaccone, R., and Caruso, G. (2002). Microbial respiratory and ectoenzymatic activities in the Northern Adriatic Sea (Mediterranean Sea). Chemistry and Ecology 18(1–2), 75–84. doi:10.1080/02757540212693 La Ferla, R., Azzaro, F., Azzaro, M., Caruso, G., Decembrini, F., Leonardi, M., Maimone, G., Monticelli, L. S., Raffa, F., Zaccone, R., and Ribera d’Alcalà, M. (2004). Microbial processes contribution to carbon biogeochemistry in the Mediterranean Sea: spatial and temporal scale variability of activities and biomass. In ‘Proceedings of ASLO/TOS 2004 Ocean Research Conference, Honolulu’. (Ed. ASLO.) p. 85. (American Society of Limnology and Oceanography: Stony Brook, NY.) Leonardi, M., and Giacobbe, S. (2001). The Oliveri-Tindari lagoon (Messina, Italy): evolution of the trophic-sedimentary environment and mollusc communities in the last twenty years. In ‘Mediterranean Ecosystems: Structures and Processes’. (Eds F. M. Faranda, L. Guglielmo and G. Spezie.) pp. 305–310. (Springer-Verlag Italia: Milano.) Nausch, M., Pollehne, F., and Kerstan, E. (1998). Extracellular enzyme activities in relation to hydrodynamics in the Pomeranian Bight (Southern Baltic Sea). Microbial Ecology 36, 251–258. doi:10.1007/S002489900112 Patel,A. B., Fukami, K., and Nishijima,T. (2000). Regulation of seasonal variability of aminopeptidase activities in surface and bottom waters of Uranouchi Inlet, Japan. Aquatic Microbial Ecology 21, 139–149. Porter, K. G., and Feig, Y. S. (1980). The use of DAPI for identifying and counting aquatic microflora. Limnology and Oceanography 25, 943–948. Puddu, A., La Ferla, R., Allegra, A., Bacci, C., Lopez, M., Oliva, F., and Pierotti, C. (1998). Seasonal and spatial distribution of bacterial production and biomass along a salinity gradient (Northern Adriatic Sea). Hydrobiologia 363, 271–282. doi:10.1023/A:1003169620843 Puddu, A., Alberighi, L., Del Negro, P., Manganelli, M., and Zaccone, R. (2000). Cicli giornalieri di produzione e abbondanza microbiche in acque costiere superficiali del Nord Adriatico. Biologia Marina Mediterranea 7, 196–205. Schultz, G. E., Jr, and Ducklow, H. (2000). Changes in bacterioplankton metabolic capabilities along a salinity gradient in the York River estuary, Virginia, USA. Aquatic Microbial Ecology 22, 163–174. Siuda, W., and Chrost, R. J. (1999). Microbial enzymatic regeneration and utilization of various organic P compounds in lake water. In ‘Proceedings of the 1st International Conference on Enzymes in the Environment: Activity, Ecology and Applications’. (Eds J. M. Barea, R. J. Chrost and R. P. Dick.) p. 77. Sommaruga, R., and Conde, D. (1997). Seasonal variability of metabolically active bacterioplankton in the euphotic zone of a hypertrophic lake. Aquatic Microbial Ecology 13, 241–248. Taylor, G. T., Way, J., Yu, Y., and Scranton, M. I. (2003). Ectohydrolase activity in surface waters of the Hudson River and western Long Island Sound estuaries. Marine Ecology Progress Series 263, 1–15. Thalassographic Reports (1990). Indagine interdisciplinare sul sistema degli stagni salmastri costieri di Oliveri-Tindari (Messina). I. Risultati delle campagne stagionali del 1987. Rapporti No 4, Consiglio Nazionale Ricerche, Istituto Sperimentale Talassografico, Messina, Italy. Thalassographic Reports (1991). Indagine interdisciplinare sul sistema degli stagni salmastri costieri di Oliveri-Tindari (Messina). II. Risultati delle campagne mensili del 1988. Rapporti No 5, Consiglio Nazionale Ricerche, Istituto Sperimentale Talassografico, Messina, Italy. Turk, V., Rehnstam, A.-S., Lundberg, E., and Hagstrom, A. (1992). Release of bacterial DNA by marine nanoflagellates, an intermediate step in phosphorus regeneration. Applied and Environmental Microbiology 58, 3744–3750.

188

Marine and Freshwater Research

G. Caruso et al.

Zaccone, R., and Caruso, G. (2002). Microbial hydrolysis of polysaccharides and organic phosphates in the NorthernAdriatic Sea. Chemistry and Ecology 18(1–2), 85–94. doi:10.1080/02757540212691 Zaccone, R., Caruso, G., La Ferla, R., Maimone, G., Puglisi, A., and Leonardi, M. (2000). Ciclo annuale dell’attività microbica nell’ecosistema lagunare di Tindari (Messina). Biologia Marina Mediterranea 7, 291–295.

Zaccone, R., Caruso, G., and Calì, C. (2002). Heterotrophic bacteria in the northern Adriatic Sea: seasonal changes and ectoenzyme profile. Marine Environmental Research 54(1), 1–19. doi:10.1016/ S0141-1136(02)00089-2 Manuscript received 17 March 2004; revised 7 February 2005; and accepted 8 February 2005.

http://www.publish.csiro.au/journals/mfr