Cultured phototrophic biofilms for phosphorus removal in wastewater treatment

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 3 5 7 – 4 3 6 7

Avai lab le a t www.sc iencedi rec t .com

journa l homepage : www.e lsev ie r . com/ loca te /wat res

Cultured phototrophic biofilms for phosphorus removal inwastewater treatment

Antonella Guzzona, Andreas Bohnb, Marco Diociaiutic, Patrizia Albertanoa,*aDepartment of Biology, University of Rome ‘‘Tor Vergata’’, Via della Ricerca scientifica 1, 00173 Rome, ItalybInstituto de Tecnologia Quımica e Biologica, Universidade Nova de Lisboa, Avenida da Republica - EAN, 2780-157 Oeiras, PortugalcDipartimento di Tecnologie e Salute, Istituto Superiore di Sanita, Viale Regina Elena 299, 00161 Rome, Italy

a r t i c l e i n f o

Article history:

Received 10 December 2007

Received in revised form

6 June 2008

Accepted 22 July 2008

Available online 3 August 2008

Keywords:

Phototrophic biofilms

Biomass production

Phosphorus removal

Wastewater

Energy Filtered-Transmission

Electron Microscopy (EF-TEM)

Electron Spectroscopy Imaging (ESI)

* Corresponding author. Fax: þ39 06 2023500E-mail address: [email protected] (P

0043-1354/$ – see front matter ª 2008 Elsevidoi:10.1016/j.watres.2008.07.029

a b s t r a c t

Culture experiments with phototrophic biofilms taken from the sedimentation tank of the

wastewater treatment plant at the Fiumicino Airport in Rome, Italy were carried out in

a prototype continuous flow incubator. Biofilms grown at varying photosynthetic photon

flux density (PPFD), temperature and flow velocity were sampled at three developmental

stages to quantify biofilm dry weight, chlorophyll a concentration and total cellular

phosphorus content. While no coherent significant effects by flow and temperature were

evidenced, maximum biofilm dry weight and phosphorous concentration significantly

increased across all featured PPFDs. Maximum chlorophyll a concentration was saturated

above 60 mmol m�2 s�1. A highly significant association between organic biomass and

phosphorous content was observed for most light conditions, including a larger propor-

tional increase of phosphorus concentration with respect to chlorophyll a at high PPFD. Up

to 112 mg P m�2 d�1 maximal phosphorous removal rates were achieved. Elemental anal-

ysis by energy filtering transmission electron microscopy showed subcellular localization

of phosphorus, confirming the accumulation in phototrophic microorganisms in biofilms

grown in high light conditions.

ª 2008 Elsevier Ltd. All rights reserved.

1. Introduction One solution to this problem is to capitalize on the nutrient

Wastewater treatment provides optimal conditions for

bacterial growth and decomposition of organic material by

water mixing and oxygenation (Gerba, 2000); however, it has

little capacity to remove inorganic nutrients (Davis et al.,

1990a). Among these, phosphorus (P) is the element normally

responsible for eutrophication of freshwater ecosystems

(Adey et al., 1993). Therefore, additional physical and/or

chemical tertiary wastewater treatment is needed before

release back into natural watercourses, but often too costly to

be implemented (Graham and Wilcox, 2000).

.. Albertano).er Ltd. All rights reserved

removal capacity of algae as a cost effective and environ-

mentally sound option for tertiary treatment of effluents

(Hoffmann, 1998; Oswald, 2003). Phosphorus is vital for algal

growth and stored when it is in excess (Graham and Wilcox,

2000). Efforts to integrate algal cultures in biological waste-

water treatment have been increasing over the last few

decades (Oswald, 1988; De la Noue et al., 1992; Chevalier et al.,

2000; Martınez et al., 2000; Kebede-Westhead et al., 2003).

Hoffmann (1998) reviewed the use of suspended algae (mixed

or monocultures) for wastewater purification in high rate algal

ponds and pinpointed the challenges involved with the

.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 3 5 7 – 4 3 6 74358

harvesting of the biomass as one of the major drawbacks of

that system.

An alternative approach to suspended algae is the use of

attached algal communities such as phototrophic biofilms

that naturally colonise the surfaces of ponds in wastewater

treatment plants (WWTPs), thus leading to easier biomass

harvesting (Hoffmann, 1998). In addition, they also exhibit

higher growth rates and therefore improved nutrient removal

capabilities (Chevalier et al., 2000). Phototrophic biofilms are

multilayered consortia of photoautotrophs (cyanobacteria

and microalgae) and heterotrophs (bacteria, fungi and

protozoa). Their layers are embedded in a common exopoly-

meric matrix, secreted by the community, that mediates the

adhesion of photo and heterotrophs, as well as gas and

nutrients fluxes (Sutherland, 2001). These communities are

ubiquitous in any aquatic environment and fulfil numerous

important environmental functions (Battin et al., 2003),

among them nutrient cycling (Wetzel, 2001), and play a key

role in the self-purification of aquatic ecosystems (Sabater

et al., 2002).

Nevertheless, the application of phototrophic biofilms in

the tertiary step of wastewater remediation is still uncommon

(Hoffmann, 1998). Previously, benthic algae were grown in

algal turf scrubber units (ATS) to remove nutrients from dairy

manure (Kebede-Westhead et al., 2003), agricultural runoff

(Adey et al., 1993) and domestic sewage (Craggs et al., 1996).

The effectiveness of ATS depends on the optimisation of the

conditions that primarily limit growth and nutrient removal

potential of biofilms, such as light, temperature and water

current (Sabater et al., 2002; Kebede-Westhead et al., 2003).

In the present work, we aim to assess the nutrient retention

capacity of phototrophic biofilms using a laboratory photo-

bioreactor specifically designed for the cultivation of

phototrophic biofilms. This system is convenient to test the

effectiveness of nutrient removal by phototrophic biofilms,

prior to field application, because it allows the simultaneous

control of the external parameters and therefore to mimic the

seasonal variations of environmental conditions occurring in

outdoor wastewater treatment plants. In particular, we

quantify the phosphorus status and maximal phosphorous

removal rate and their correlation with the phototrophic

biomass production of biofilms sampled from an Italian

WWTP and grown in cultures at varying irradiance, tempera-

ture and medium flow rate. As P uptake at community level

crucially depends on the diversity of nutrient storage capacity

of the algal species in the biofilm (Sabater et al., 2002), we

also address the relative contribution of the different

phototrophic components to P retention using Energy Filtered-

Transmission Electron Microscopy (EF-TEM) and Electron

Spectroscopy Imaging (ESI) techniques to localize cellular P

reserves of the phototrophic species dominating the biofilm.

2. Materials and methods

2.1. Study site

The sample site was a sedimentation tank (ST) of the waste-

water treatment plant (WWTP) that serves the ‘‘Leonardo da

Vinci’’ Airport in Fiumicino (Rome, Italy) (41�470N, 12�150E).

The plant is an outdoor system and receives domestic sewage

at an estimated through-flow of 6000 m3 d�1. Water treatment

involves oxygen injections in the primary tank and constant

stirring in the secondary tank to reduce biochemical oxygen

demand (B.O.D.). The ST is located above ground and is fully

exposed to sunlight. Standard water treatment procedures

involved weekly scrubbing of tank walls to avoid excessive

biomass accumulation.

Due to the constant water flow and the high concentra-

tions of inorganic nutrients, the Fiumicino WWTP could be

classified as an artificial hypertrophic lotic system (OECD,

1982). In situ variables have been previously reported by the

authors, in brief; temperature, 15–25 �C; irradiance, 690–

1550 mmol m�2 s�1; pH, 6.5–7.7; TP, 0.6–3.0 mg l�1; NO3, 0.2–

8.2 mg l�1; NH4, 1.6–20 mg l�1 (Albertano et al., 1999; Guzzon

and Albertano, in press).

2.2. Biofilm sampling

Phototrophic biofilm samples were taken from the sedimen-

tation tank (ST) on two occasions in both spring 2004 (April,

May) and summer 2004 (July and September). Around 70 g

(fresh weight) of biofilm were scraped off a vertical ledge in the

ST walls the day before the weekly tank cleaning. The samples

were placed in plastic bottles, stored in a cool box with ice and

transported to the laboratory. Before the samples were used

they were pre-treated to obtain a safe-to-handle and homo-

geneous inoculum. To do this the samples were washed in 2%

sodium hypochlorite for 15 min to kill zoobenthos and path-

ogens and then rinsed in a modified BG11 medium (Stanier

et al., 1971). Modifications included adding vitamins and sili-

cates to support microalgal growth (40 mg l�1 and 57 mg l�1,

respectively, according to Beakes et al., 1988), using only

nitrate (NO3�) as the sole nitrogen source and increasing the

concentration of orthophosphate (PO43�) to obtain a final

nitrogen:phosphorus (N:P) ratio of 15 (P 3.64 mg l�1), thus

ensuring balanced conditions for algal growth. Thereafter,

biofilm biomass was homogenised using a blender, allowed to

settle for 1 h and the final suspension was frozen in fresh

medium (�20 �C for 24 h) to remove grazers (Guzzon et al.,

2005).

2.3. Setup of biofilm cultures

Biofilms were cultured in a prototype phototrophic biofilm

incubator (PBI, Zippel and Neu, 2005; Zippel et al., 2007) (Fig. 1)

consisting of four individual flow lanes each measuring

120� 10 cm. In each lane, medium flowed over polycarbonate

slides (76� 25� 1 mm) which were used as substratum for

biofilm adhesion (Fig. 1a). A turbulence reducer and a water

temperature sensor were located at the inlet device of each

chamber and flow rate was valve regulated. Each lane had

a specific light intensity. Light was provided by fluorescent

lamps (True-light 36 W, Auralight, Sweden), connected to

a timer to produce a 16:8 h light–dark cycle. Photosynthetic

photon flux density (PPFD, mmol m�2 s�1) was measured by an

incident light sensor placed under each lamp (Fig. 1b). Biofilm

growth was monitored by three subsurface light sensors

located directly under three selected slides in each flow lane

(Fig. 1c and d).

Fig. 1 – Prototype of the incubator used to culture the phototrophic biofilms (a). Inside view of the lane showing the slides used

as substratum and the lamp positioned above the chamber (b). The surface sensor was located under the lamp while

subsurface sensor positioned under the slides (c). Mature biofilms colonising the slides (d).

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 3 5 7 – 4 3 6 7 4359

Incubation involved diluting 100 mL of the homogenised

treated biofilm inoculum in modified BG11 medium to a final

volume of 400 mL. Four aliquots of 100 mL were then added to

3.9 L of medium and filtered through a 300 mm net. This

solution was then circulated through the incubator for 72 h at

0.15 m s�1. Thereafter, this solution was replaced with fresh

medium and the flow rate was set at 5 or 15 cm s�1 (Table 1).

The cultures were kept for 3 days under nutrient replete

conditions by changing the medium twice a week.

Biofilms were cultured at 120 (LC120), 60 (LC60), 30 (LC30)

and 15 mmol m�2 s�1 (LC15) (Table 1). At each PPFD, two

Table 1 – Sampling month, temperatures and mediumflow rates in the incubator runs performed

Run Label Sampling month(year 2004)

Temperature[�C]

Flow rate[cm s�1]

1 R 20.05 April 20 5

2 R 20.15 June 20 15

3 R 30.05 July 30 5

4 R 30.15 September 30 15

temperatures and medium flow rates were tested resulting in

four experimental runs (Table 1). Each run lasted 30 days.

Biofilms were scraped off the slides at three stages of devel-

opment given as a function of incident light absorption by the

biofilm: (1) initial stage, when the absorption A, averaged over

the three sensors in each lane, equalled 10%; (2) active stage,

when A¼ 50%; (3) mature stage, at A¼ 90–95%. On the last day

of the run biofilm was collected even if the active or mature

stage had not been reached.

2.4. Quantification of biofilm biomass and totalphosphorus (TP)

Biofilm biomass (phototrophsþheterotrophs) was deter-

mined as dry weight (DW) by oven drying samples at 60 �C for

72 h. Dried biomass was then burnt for 1 h at 500 �C to deter-

mine the organic biomass as AFDW (APHA, 1998). Photo-

trophic biomass was assessed by determining chlorophyll

a concentration per unit area (ChlA). Chlorophyll a was

extracted overnight in 90% acetone in the dark and then

quantified spectrophotometrically according to Jeffrey and

Humphrey (1975) (in Jensen, 1978). Ashed biomass was

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 3 5 7 – 4 3 6 74360

analysed for TP by means of the ascorbic acid method (APHA,

1998). All analyses of biomasses and TP were performed in

triplicate.

2.5. Estimation of the maximal phosphorousremoval rate

The maximal removal rate for phosphorous (MRP) per surface

unit between two time points t1< t2 was estimated by multi-

plying the production rate (dry weight, DW) with the average

percentual phosphorous content, according to Davis et al.

(1990b).

MRPt1/t2¼ DW2 �DW1

t2 � t1

PCont1 þ PCont2

2

2.6. Electron microscopy observations

Biofilm samples were fixed in 2.5% glutaraldehyde in caco-

dylate buffer 0.2 M pH 7.2, post-fixed in 1% osmium tetraoxide

in the same buffer, dehydrated in ethanol series and

embedded in an 812 Resin� Kit. Thin sections of 60 nm were

assembled on copper grids of 300 mesh, double stained with

2% uranyl acetate and 1% lead citrate (Reynolds, 1963) and

observed by a Zeiss EM 902 electron microscope operating at

80 keV, equipped with an electron energy loss filter at 80 kV in

conventional mode. Species composition of the phototrophic

biofilms colonising the ST of Fiumicino WWTP has been

previously reported (Albertano et al., 1999; Congestri et al.,

2003, 2005).

For Electron Spectroscopy Imaging (ESI), ultrathin sections

(<40 nm) were placed on 600-mesh copper grids and observed

without staining. Images were acquired filtering the elastic

electrons (DE¼ 0 eV) to enhance both contrast and resolution

(Diociaiuti, 2005). The image acquisition was performed by

a digital CCD camera model HSC2, 1 k for 1 k pixels (Proscan

GMBH, Germany), thermostatted by a Peltier cooler model

WKL 230 (LAUDA GMBH, Germany). Image analysis and

quantification were performed by a digital analyser analysis

3.0 (SIS GMBH, Germany). Subcellular localization of P in the

specimen was obtained according to the so-called ‘‘three-

window methods’’ (Egerton, 1986). Briefly, a reference image

with high contrast was acquired filtering at 250 eV. In order to

obtain the P net distribution, after the acquisition of two pre-

edge images (at 110 and 130 eV) the software calculated the

intensity of the background image to be subtracted from the

post-edge image (at 153 eV). Finally, the net image was con-

verted to a binary image by a grey-level discrimination

threshold.

2.7. Statistical analyses

Statistical analyses were carried out using the R language,

version 2.6.2 for Mac OS X (R Development Core Team, 2008)

and Aabel software, version 2.4.2 (Gigawiz Ltd. Co., Tulsa, OK)

for graphical representations. Log-transformed data were

used where necessary. Correlations were quantified by Pear-

son’s r. The effects of the factors run (temperature and flow

velocity) and lane (light intensity) on the 30-day maxima of

PConc, AFDW, and ChlA were assessed by two-way ANOVA and

post-hoc testing for differences of means using Tukey’s

Honest Significant Difference (Tukey’s HSD) method (Yandell,

1997). Elasticity analysis (De Kroon et al., 2000) was applied to

compare the percent changes of two variables: given a relation

between the two variables x and y, the elasticity e, i.e. the ratio

of the percent changes of x and y, can be estimated by the

slope of the linear regression of log( y) vs. log(x).

e ¼ dy=ydx=x

¼ d logðyÞd logðxÞ

If e> 1 the proportional change of y is larger than the cor-

responding proportional change in x, and vice versa for e< 1.

In the present case, we interpret e> 1 as a relative accumu-

lation of y in comparison to x. Values around e¼ 1 indicate

similar proportional growth. Computationally, e was esti-

mated by standardised major axis estimation (Warton et al.,

2006), using the R package smatr (Warton and Ormerod, 2007),

which includes statistical testing of the null hypothesis of

e¼ 1.

3. Results

3.1. Biomass accumulation, phosphorus contents andremoval rate

Dry weight (DW) of the sampled biofilms was comprised

between 0.5� 0.2 and 87.1� 11.1 g m�2 (Fig. 2A–D). Without

considering all details, the visual inspection of the time series

reveals a monotonous increase with time in all lanes and for

all runs. The same monotonous development in time is

observed for the total P concentration (PConc, Fig. 2E–H), which

shows values between 3.5� 0.3 and 666� 67 mg m�2. An

exception is the sample LC30 grown at 20 �C and 5 cm s�1 (R

20.05), which shows very high values in comparison to the

other runs. However, the large error margins of these values

do not indicate a high significance of this development. No

coherent temporal development can be observed in the total

phosphorous content (PCont) in relation to the biofilm DW

(%w/w) (Fig. 2I–L). It shows a range of values between 0.1� 0.1

and 6.9� 0.9% w/w. The maximal areal removal rates for

phosphorus (MRP) cover a span between 0.3� 0.3 and

111.9� 14.2 mg P m�2 d�1 (Fig. 2M–P).

Even though the dynamics of the directly observable vari-

able PConc and the technically relevant MRP do not show

a clear correlation, we can show that their maximum values

attained during the 30 days of observation are strongly

correlated (Fig. 3). Correlation analysis yields r¼ 0.89,

p< 0.001. PConc can thus be considered as a good indicator for

MRP.

In Fig. 4 we examine the correlation between PConc and DW,

ash-free dry weight (AFDW), and chlorophyll a concentration

(ChlA). The details of the temporal dynamics of the two latter

quantities will be discussed in a forthcoming work and are

thus not explicitly shown here. Visual inspection of the entire

sets of points in Fig. 3 and quantitative correlation analysis

(Table 2) indicate that PConc is most strongly related to AFDW

(r¼ 0.92), ahead of DW (r¼ 0.88) and ChlA (r¼ 0.73). The

analysis discriminated by lane shows that a significant asso-

ciation of PConc and AFDW persists in all light conditions,

0

250

500

750

PC

on

c [m

g m

-2]

0

2

4

6

8

PC

on

t [%

w

/w

]

0

25

50

75

100D

W [g

m

-2]

DayDayDay

7 14 21 28 35 7 14 21 28 35 7 14 21 28 35 7 14 21 28 35Day

0

50

100

150

MR

P [m

g m

-2 d

-1]

A B C D

E F G

I J K

M N O P

LC 15 LC 30 LC 60 LC 120

H

L

Fig. 2 – Temporal development of biomass and phosphorous parameters. Rows (from top): dry weight (A–D), total P

concentration (E–H), specific P contents (I–L), maximum P removal rate (M–P). Columns (from the left): light chambers LC15

through LC120. Symbols: ,, R 20.05; 6, R 20.15; D, R 30.05; 3, R 30.15.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 3 5 7 – 4 3 6 7 4361

while a significant correlation of PConc and both DW and ChlA

is only given with exception of the lowest light intensity.

Comparing the relative growth rates by means of elasticities,

it shows that both DW and AFDW grow at rates similar to

PConc, as in no case the elasticity diverges significantly from

e¼ 1 (Table 2). This is different for the case of ChlA: consid-

ering all points, the elasticity is e¼ 0.7, indicating a higher

proportional growth of ChlA with respect to PConc. However,

for the highest light intensity (LC120), e¼ 1.75 holds, indi-

cating a higher proportional growth, i.e. accumulation, of

PConc with respect to ChlA, at least to marginal significance

( p¼ 0.054).

3.2. Environmental effects on 30-day maxima

To further elucidate the findings of the previous correlation

analysis and to assess environmental effects, the 30-day

maxima of PConc, AFDW, and ChlA against PPFD are shown in

Fig. 5. Given the strong correlation of AFDW with PConc and

DW (data not shown), for this part of our study we focus on

AFDW as a measure of biomass. Within reasonable limits, the

results and conclusions drawn also apply to total biofilm dry

weight.

PConc Max and AFDWMax increase across the entire range of

PPFDs, while ChlAMax is saturated above PFD¼ 60 mE m�2 s�1,

reaching a level of about 125 mg m�2. This is confirmed by

quantitative analyses (Table 3), which shows that for PConc and

AFDW the increases from lane to lane are all significant, with

the exception of AFDW between LC30 and LC15. Testing for

differences between different light intensities for ChlA yields

significance at p¼ 0.05 in all cases, except the difference

between LC120 and LC60, confirming the saturation effect

observed in Fig. 5.

Assessing the factor run (temperature and flow velocity),

the two-way ANOVA indicates a significant effect for both

PConc and AFDW. The former observable features a significant

interaction between runs and lanes, i.e. the differences

induced by temperature and flow velocity are dependent on

the light intensity. From Tukey’s HSD-testing and Fig. 5A we

can infer that R 20.15 (low temperature, high flow) shows

significantly lower values than all other runs. This would

indicate a positive effect of low temperature, and a negative

effect of high temperature and low flow velocity on the

0.5 1.0 1.5 2.0 2.5 3.0Log

10 P

Conc Max

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

Lo

g1

0 M

RP

Max

Fig. 3 – Log–log scatterplot of 30-day maximum P removal

rate vs. 30-day maxima of P concentration. Symbols: ,,

LC120; 6, LC60; D, LC30; 3, LC15.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 3 5 7 – 4 3 6 74362

PConc Max. On the other hand, the values in R30.05 are signifi-

cantly smaller than in R 20.05 and R 30.15, contradicting these

hypotheses. We thus cannot find a significant and coherent

effect of either temperature or flow velocity on maximal

phosphorus concentration.

Verifying the effect of temperature and flow velocity on the

maximum AFDW values, R 20.05 in LC30 yielded an outlying

point, which caused a significant interaction between lanes

and runs. Run 20.05 was therefore excluded from the analysis.

For the remaining runs, the Tukey HSD test indicates

a significantly larger maximum AFDW at 30 �C (R 30.15) in

comparison to 20 �C (R 30.15) in low flow conditions, indicating

a positive effect of high temperature on the accumulation of

AFDW. Given our contradictory data, this remains a weak

hypothesis needing further experimental examination.

3.3. TEM and ESI observations

Ultrastructural observations were performed on samples

grown at 120 mmol m�2 s�1 as they were the only cultures that

developed into mature biofilms. High light biofilms were

mainly composed of cyanobacteria, both filamentous, such as

Phormidium sp. (Fig. 6a) and Pseudanabaena sp., and unicellular

forms, such as Synechocystis sp. In Phormidium sp. trichomes

there were several cyanophycin granules spread throughout

the cytoplasm or localized near to the cross walls (Fig. 6c).

-0.5Log

10 DW

0.5

1.0

1.5

2.0

2.5

3.0

Lo

g10 P

Co

nc

Log1

0.0 0.5 1.0 1.5 2.0 -1.0 -0.5 0.0

A

Fig. 4 – Relation of P concentration with biomass. (A) Log–log scat

,, LC120; 6, LC60; D, LC30; 3, LC15.

Coccal unicellular green algae , such as Chlorococcum sp., and

colonial, such as Scenedesmus sp. and Desmodesmus sp. (Fig. 6b

and d), also occurred. In green algae, starch granules were

present, both around the pyrenoid and between thylakoids

(Fig. 6a and d). Partly metabolised polyphosphates were visible

in the cells of both Chlorococcum sp. and Scenedesmus sp.

colonies (Fig. 6a). Numerous polyphosphate bodies (diameter

up to 0.5 mm in Chlorococcum sp. and Scenedesmus sp., and up to

0.8 mm in Desmodesmus sp.) occurred in the vacuoles of the

green algae (Fig. 6b and d).

In Fig. 7 we present the maps showing the net distribution

of phosphorus inside the cells of phototrophs obtained by ESI.

To give an example for this technique, the high contrast image

(DE¼ 250 eV) and the P net distribution used to obtain the map

of P distribution are shown in Fig. 7a and b for a cell of the

green alga Desmodesmus sp. At DE¼ 250 eV it was possible to

clearly distinguish the cell wall, from which a short spina is

protruding, the chloroplast and thylakoidal membranes,

along with the nucleus and organelles in cytoplasm. The map

showing the net P distribution (Fig. 7c), obtained after super-

imposing the net distribution of the element (Fig. 7b) on the

high contrast image (Fig. 7a), showed that P was localized

mainly in the dark starch granules occurring between the

thylakoids and in the white zone observed in the nuclear area

at DE¼ 250 eV. In addition, ESI analyses of other cells of the

green alga Desmodesmus showed that P was also localized in

the starch around the pyrenoid and in the cortical cytoplasm

(Fig. 7d) as well as in the numerous polyphosphate bodies that

accumulated in the vacuoles of cells at the mature stage

(Fig. 7e), as previously observed in conventional mode (Fig. 6d).

As for cyanobacteria, P was localized between thylakoids and

spread through the cytoplasm of the Phormidium sp. (Fig. 7f).

4. Discussion

We performed culture experiments to quantify biomass

production and nutrient status of phototrophic biofilms,

assessing the influence of the environmental factors such as

light intensity, flow velocity and temperature in a flow lane

incubator. In addition to quantitative analyses of biofilm

samples from different stages of biofilm development, phos-

phorous accumulation in phototrophs was assessed with TEM

and ESI microscopy.

0 AFDW

LC 120LC 60LC 30LC 15

0.5 1.0 1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5Log

10 ChlA

B C

terplot of PConc vs. DW (A), AFDW (B), and ChlA (C). Symbols:

Table 2 – Results of correlation and elasticity analyses between phosphorous and DW, AFDW, and ChlA, respectively

DW AFDW ChlA

r p (r¼ 0) e p (e¼ 1) r p (r¼ 0) e p (e¼ 1) r p (r¼ 0) e p (e¼ 1)

All 0.88 <0.001 0.92 0.300 0.92 <0.001 0.99 >0.5 0.73 <0.001 0.70 0.003

LC120 0.83 0.001 1.16 0.444 0.86 0.001 1.01 >0.5 0.59 0.052 1.75 0.054

LC60 0.91 <0.001 0.86 0.288 0.94 <0.001 1.00 >0.5 0.80 0.002 0.84 0.392

LC30 0.91 0.001 1.14 0.406 0.94 <0.001 0.97 >0.5 0.82 0.006 0.90 >0.5

LC15 0.79 0.060 1.91 0.085 0.90 0.014 1.23 0.386 0.34 >0.5 1.11 >0.5

Italicized numbers indicate significance at p¼ 0.05 level.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 3 5 7 – 4 3 6 7 4363

A technically important result is the suitability of the

featured flow culture system for the reproduction of biofilms

with high phototrophic biomass, starting from a natural

sample whose structure had been experimentally altered to

obtain a homogenous and grazers-free cultivable inoculum.

Chlorophyll a concentrations showed in fact that photo-

trophic biomass production in cultured biofilms after 30 days

was in the range obtained for the same biofilms after 7 days

growth in situ (4.3–252.6 mg m�2, from Guzzon and Albertano,

in press). P content attained maximum values in the range of

the natural community from the same plant (6.9% w/w vs.

3.5% from Albertano et al., 2003). Also, maximum P removal

10-1

100

101

AF

DW

[m

g m

-2]

0 30 60 90 120 150PPFD [µE m

-2 s

-1]

10-1

100

101

102

Ch

lA

[µ

g m

-2]

100

101

102

103

PC

on

c [m

g m

-2]

A

B

C

Fig. 5 – 30-Day maximum values of P concentration and

biomasses vs. light intensity. (A) PConc, (B) AFDW, (C) ChlA.

Symbols: ,, R 20.05; 6, R 20.15; D, R 30.05, 3 R 30.15.

rates were comparable to those obtained for ST biofilms from

an American wastewater treatment plant grown in situ (Davis

et al., 1990b).

The most important environmental factor for biofilm

development in the culture system used is photosynthetic

photon flux density (PPFD). The maximum values of total and

organic biomass (DW and AFDW), as well as PConc, increase

significantly throughout the entire range of PPFDs used here

(Fig. 4). This is expected because light promotes the growth of

phototrophic microorganisms (Hill, 1996) and P is required for

cell biosynthesis. The strong correlation of PConc and AFDW on

one hand, and PConc Max and the maximum areal removal rate

of P (MRP) on the other, implies that, under any light condi-

tions, P is likely removed and used for the production of the

photo and heterotrophic biomass of the biofilm. This same

relationship was also found in periphyton of the Floridian

Everglades (Vymazal and Richardson, 1995).

Maximum ChlA shows a nonlinear dependency on PPFD,

saturating above 60 mmol m�2 s�1, most likely due to satura-

tion of photosynthesis (Hill, 1996; Hill and Fanta, 2007). This

complies with previous studies on the light acclimation

response of the ST biofilms, grown on artificial substrata in the

plant, which yielded values of saturating irradiance around

Table 3 – Results of two-way ANOVA and Tukey’s HSD-testing with factors run and LC for maximal values ofPConc, AFDW, ChlA (Fig 3)

PConc AFDW ChlA

Two-way ANOVA Run <0.001 0.002 0.284

Lane <0.001 <0.001 <0.001

Run:lane <0.001 0.157 0.111

Differences of runs 20.15� 20.05 <0.001 ND 0.576

30.05� 20.05 <0.001 ND 0.280

30.15� 20.05 0.103 ND 0.400

30.05� 20.15 0.005 0.213 >0.5

30.15� 20.15 <0.001 <0.001 >0.5

30.15� 30.05 0.076 0.070 >0.5

Differences of LCs 60� 120 0.014 0.043 >0.5

30� 120 <0.001 <0.001 0.035

15� 120 <0.001 <0.001 <0.001

30� 60 <0.001 <0.001 0.011

15� 60 <0.001 <0.001 <0.001

15� 30 <0.001 0.067 <0.001

Italicized numbers indicate significance at p¼ 0.05 level.

Fig. 6 – TEMmicrographsofLC120biofilmwithcross-andlongitudinalsectionsofChlorococcumsp.,Scenedesmussp.andPhormidiumsp.

(top tobottom) (a).Partly metabolisedpolyphosphates (a–white arrows)werevisible inthecellsofgreenalgae.Numerouscyanophycin

granules occurred spread in the cytoplasm or near to cross cell walls in Phormidium sp. trichomes (c – black arrows). Conspicuous

polyphosphate bodies ( p) were localized inside vacuoles within the cytoplasm of green algae (b, d). Where unspecified bar [ 1 mm.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 3 5 7 – 4 3 6 74364

60 mmol m�2 s�1 (Guzzon and Albertano, in press). A similar

saturation effect has been observed in parallel experiments

with ST biofilms grown under the same conditions in the

same flow lane incubator but located in a different laboratory

(Zippel et al., 2007).

Our statistical analyses confirm the accumulation of PConc

and organic biomass with respect to ChlA at the highest light

intensity tested. An open question, which cannot be inferred

from these data, is whether P accumulation occurs in the

photo or the heterotrophic portion of the biofilm. Neverthe-

less, our imaging efforts using TEM and ESI show that P

reserves occur inside the cells of the phototrophs of high light

cultures. In particular, the green algae seemed to be accu-

mulating large amounts of P since polyphosphate storages

were observed in the cytoplasm of the green algae Chlor-

ococcum sp., Scenedesmus sp. and Desmodesmus sp. (Fig. 6).

Among these species, Scenedesmus sp. is already known for

storing P and has already been employed as monoculture in

wastewater depuration from P (Martınez et al., 2000). From our

observations, also Desmodesmus sp. could be a suitable species

to be implied in the nutrient removal from wastewaters since

numerous P storages were revealed by ESI analyses. ESI

images showed that, in this species, the level of P storage

increased from small reserves between thylakoids in the

active stage of growth to numerous and enlarged inclusion

bodies in the vacuole in mature cultures, evidencing P accu-

mulation in later phase of biofilm development.

Finally, the effects of flow velocity and temperature did not

evidence a clear effect of either factor on P retention by pho-

totrophic biofilms. The most likely reason is that sample

variability overrode the effects induced by the applied changes

of the environmental factors. The hypothesis that the chosen

temperature range was insufficient to yield a significant effect

is sustained by Defew et al. (2004), who found significant

effects of temperature on cultured microphytobenthic

assemblages when the temperature was increased of 8 �C

from the field, whilst in this study the highest temperature

was only 5 �C higher than the maximum measured in the

Fig. 7 – ESI images of the phototrophs from high light biofilms. High contrast image (DE [ 250 eV) of Desmodesmus sp. (a). Net

phosphorus distribution obtained by the subtraction of two pre-edge images (DE [ 110 and 130 eV) from the post-edge image

(DE [ 153 eV) (b). Elemental mapping resulting from the superimposition of images (a) and (b), showing the localization of P in

the dark starch granules between thylakoids and in the white vacuolar central area. P mapping in the cyanobacterium

Phormidium sp. (d) and in the green alga Desmodesmus sp. in biofilms at the active (e) and mature (f) phase of development.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 3 5 7 – 4 3 6 7 4365

sedimentation tank (Guzzon and Albertano, in press). Future

experimental designs should thus take particular care in

reducing the natural sample variability and choosing suffi-

ciently discriminating parameter sets.

For the time being, we may conclude that the storage

capacity for P of phototrophic organisms turns cultured pho-

totrophic biofilms into promising candidates for P removal in

wastewater treatment. Future biotechnological work might

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 3 5 7 – 4 3 6 74366

depart from our hypothesis that high light intensity is the

most crucial factor for achieving high P retention rates, and

elucidate the precise effects of temperature and flow velocity

on nutrient retention by phototrophic biofilms.

5. Conclusions

� Biofilm dry weight and total P concentration increase with

the increase of the featured light intensities. Maximum

chlorophyll a concentration increases up to 60 mmol m�2 s�1

and is saturated above this level. Optimal illumination and

avoidance of saturation of photosynthesis are thus essential

for the optimized application of phototrophic biofilms in

wastewater treatment.

� Total P concentration is strongly related with organic

biomass throughout all light conditions. The maximum P

concentration observed during an incubator run is a good

estimator for the maximum P removal rate of the biofilm.

� At the highest light intensity (120 mmol m�2 s�1) phosphorus

and organic biomass are accumulated with respect to ChlA.

TEM and ESI observations show that phosphorus is accumu-

lated insidethecellsofphototrophs,mainly inthecytoplasmof

the green algae. Coupling TEM–ESI techniques to the quantifi-

cation of nutrient removal is therefore useful to screen and

select the phototrophic speciesmost suitable to be cultivated in

sewage plants for the tertiary treatment of wastewaters.

� Effects of changes in environmental parameters, in our case

temperature and flow velocity, can be masked by natural

sample variability.

Acknowledgements

We gratefully acknowledge the Agency for Environmental

Protection of the ‘‘Aeroporti di Roma’’, Fiumicino (Italy) for the

fruitful collaboration and the granted access to the plant.

This work was supported by the EU-Programme Quality of Life,

in the frame of PHOBIA Project, contract QLK3-CT2002-01938. The

incubator prototype was realized by the Department of Inland

Water Research, UFZ Centre for Environmental Research (Mag-

deburg,Germany) intheframeworkof theEU-projectPHOBIA.A.B.

acknowledges support by the Fundacao para a Ciencia e a Tecno-

logia (Portugal) through fellowship SFRH/BPD/25967/2005. We

thank Nuno Sepulveda for fruitful discussions. Dr Neil T.W. Ell-

wood revised the manuscript providing many useful comments.

r e f e r e n c e s

Adey,W.,Luckett,C., Jensen,K., 1993.Phosphorusremoval fromnaturalwaters using controlled algal production. Restor. Ecol., 29–39.

Albertano, P., Congestri, R., Shubert, L.E., 1999. Cyanobacterialbiofilms in sewage treatment plants along the Tyrrheniancoast (Mediterranean Sea), Italy. Algol. Stud. 94, 13–24.

Albertano, P., Guzzon, A., Di Pippo, F., Congestri R., 2003. Potentialuse of phototrophic biofilms for N and P removal. In: Book ofAbstracts of the 5th European Workshop ‘‘Biotechnology ofMicroalgae’’, Bergholz-Renbrucke, Germany, June 23–24.

APHA, 1998. Standard Methods for the Examination of Waterand Wastewater. In: Clesceri, L.S., Geenberg, A.E., Eaton, A.D. (Eds.), 20th ed. APHA, AWWA, WEF, Washington, D.C.

Battin, T.J., Kaplan, L.A., Newbold, J.D., Hansen, C.M.E., 2003.Contributions of microbial biofilms to ecosystem processes instream mesocosm. Nature 426, 439–442.

Beakes, G., Canter, H.M., Jaworski, G.H.M., 1988. Zoosporesultrastructure of Zygorhizidium affluens Canter and Z.planktonicum Canter, two chytrids parasitizing the diatomAsterionella formosa Hassall. Can. J. Bot. 66 (6), 1054–1067.

Chevalier, P., Proulx, D., Lessard, P., Vincent, W.F., de la Noue, J.,2000. Nitrogen and phosphorus removal by high latitude mat-forming cyanobacteria for potential use in tertiary wastewatertreatment. J. Appl. Phycol. 12, 105–112.

Congestri, R., Sangiorgi, V.C., Albertano, P., 2003. Cytomorphologyand distribution of periphytic cyanobacteria in one ItalianWWTP. Arch. Hydrobiol., Algol. Stud. 109, 185–195.

Congestri, R., Cox, J.E., Cavacini, P., Albertano, P., 2005. Diatoms(Bacillariophyta) in phototrophic biofilms colonising anItalian wastewater treatment plant. Diatom Res. 20 (2),241–255.

Craggs, R.J., Adey, W.H., Jenson, K.R., St. John, M.S., Green, F.B.,Oswald, W.J., 1996. Phosphorus removal from wastewater usingan algal turf scrubber. Water Sci. Technol. 33 (7), 191–198.

Davis, L.S., Hoffmann, J.P., Cook, P.W., 1990a. Seasonalsuccession of algal periphyton from a wastewater treatmentfacility. J. Phycol. 26, 611–617.

Davis, L.S., Hoffmann, J.P., Cook, P.W., 1990b. Production andnutrient accumulation by periphyton in a wastewatertreatment facility. J. Phycol. 26, 617–623.

De la Noue, J., Laliberte, G., Proulx, D., 1992. Algae andwastewater. J. Appl. Phycol. 4, 247–254.

Defew, E.C., Perkins, R.G., Paterson, D.M., 2004. The influence oflight and temperature interactions on a natural estuarinemicrophytobenthic assemblage. Biofilms 1, 21–30.

De Kroon,H., van Groenendaal, J., Ehrlen, J., 2000. Elasticities: a reviewof methods and model limitations. Ecology 81 (3), 607–618.

Diociaiuti, M., 2005. Electron energy loss spectroscopymicroanalysis and imaging in the transmission electronmicroscope: example of biological applications. J. ElectronSpectrosc. Relat. Phenom. 143, 189–203.

Egerton, R.F., 1986. Electron Energy Loss Spectroscopy in theElectron Microscope. Plenum Press, New York.

Gerba, C.P., 2000. Domestic wastes and waste treatment. In:Maier, R.M., Pepper, I.L., Gerba, C.P. (Eds.), EnvironmentalMicrobiology. Academic Press, Canada, pp. 505–534.

Graham, L.E., Wilcox, L.W., 2000. Algae. In: Ryu, T. (Ed.). PrenticeHall, NJ, pp. 1–640.

Guzzon, A., Albertano, P. Annual photosynthetic response ofphototrophic biofilms from an Italian wastewater treatmentplant. Algol. Stud., in press.

Guzzon, A., Congestri, R., Albertano, P., 2005. Light-inducedchanges in photosynthesis and structure of cyanobacteriacultured biofilms from an Italian wastewater treatment plant.Arch. Hydrobiol., Algol. Stud. 117, 223–228.

Hill, W.R., 1996. Effects of light. In: Stevenson, R.J., Bothwell, M.L.,Lowe, R.L. (Eds.), Algal Ecology – Freshwater BenthicEcosystems. Academic Press, USA, pp. 121–148.

Hill, W.R., Fanta, S.E., 2007. Phosphorus and light colimitperiphyton growth at subsaturating irradiances. Freshw. Biol.doi:10.1111/j.1365-2427.2007.01885.x.

Hoffmann, J.P., 1998. Wastewater treatment with suspended andnonsuspended algae. J. Phycol. 34, 757–763.

Jensen, A., 1978. Physiological & biochemical methods. In:Hellebust, J.A., Craigie, J.S. (Eds.), Handbook of PhycologicalMethods. Cambridge University Press, Cambridge.

Kebede-Westhead, E., Pizarro, C., Mulbry, W.W., 2003. Productionand nutrient removal by periphyton grown under different

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 4 3 5 7 – 4 3 6 7 4367

loading rates of anaerobically manure digested flushed dairy.J. Phycol. 39, 1275–1282.

Martınez, M.E., Sanchez, S., Jimenez, J.M., El Yousfi, F., Munoz, L.,2000. Nitrogen and phosphorus removal from urbanwastewater by the microalga Scenedesmus obliquus. Bioresour.Technol. 73, 263–272.

Organisation of Economic Co-operation and Development,1982. Eutrophication of Waters: Monitoring andAssessment. OECD, Paris.

Oswald, W.J., 1988. Micro-algae and waste-water treatment. In:Borowitzka, M.A., Borowitzka, L.J. (Eds.), MicroalgalBiotechnology. Cambridge University Press, Cambridge, pp.305–328.

Oswald, W.J., 2003. My sixty years in applied algology. J. Appl.Phycol. 15, 99–106.

Reynolds, C.S., 1963. The use of lead citrate at high pH as anelectronopaque stain in electron microscopy. J. Cell. Biol. 17,208–212.

R Development Core Team, 2008. R: a Language and Environmentfor Statistical Computing. R Foundation for StatisticalComputing, Vienna, Austria. Available from: http://www.R-project.org.

Sabater, S., Guasch, H., Romanı, A., Munoz, I., 2002. The effect ofbiological factors on the efficiency of river biofilms inimproving water quality. Hydrobiologia 469, 149–156.

Stanier, R.Y., Kunisawa, R., Mandel, M., Cohen-Bazire, G., 1971.Purification and properties of unicellular blue-green algae(Order Chroococcales). Bacteriol. Rev. 35, 171–205.

Sutherland, I.W., 2001. Biofilm exopolysaccharides: a strong andsticky framework. Microbiology 147, 3–9.

Vymazal, J., Richardson, C.J., 1995. Species composition, biomassand nutrient content of periphyton in the Florida Everglades.J. Phycol. 31, 343–354.

Warton, D.I., Wright, I.J., Falster, D.S., Westoby, M., 2006. Bivariateline-fitting methods for allometry. Biol. Rev. 81, 259–291.

Warton, D.I., Ormerod, J., 2007. smatr: (Standardised) Major AxisEstimation and Testing Routines. R Package Version 2.1.Available from: http://web.maths.unsw.edu.au/wdwarton.

Wetzel, R.G., 2001. Land–water interfaces: attachedmicroorganisms, littoral algae, and zooplankton. In:Limnology: Lake and River Ecosystems. Academic Press, SanDiego, pp. 577–623.

Yandell, B.S., 1997. Practical Data Analysis for DesignedExperiments. Chapman & Hall.

Zippel, B., Neu, T.R., 2005. Growth and structure of phototrophicbiofilms under controlled light conditions. Water Sci. Technol.52 (7), 203–209.

Zippel, B., Rijstenbil, J., Neu, T.R., 2007. A flow-lane incubatorfor studying freshwater and marine phototrophic biofilms.J. Microbiol. Methods 70, 336–345.