Ecological Engineering 82 (2015) 49–56

Bacteria viability and decay in water and soil of vertical subsurface flowconstructed wetlands

P. Foladori a,*, L. Bruni b, S. Tamburini c

aDepartment of Civil, Environmental and Mechanical Engineering, University of Trento, via Mesiano 77, 38123 Trento, ItalybAgenzia per la Depurazione, Autonomous Province of Trento, via Gilli 3, 38122 Trento, ItalycCentre for Integrative Biology (CIBIO), University of Trento, Via delle Regole 101, Mattarello, 38123 Trento, Italy

A R T I C L E I N F O

Article history:Received 29 April 2014Received in revised form 4 March 2015Accepted 6 April 2015Available online xxx

Keywords:WastewaterVertical subsurface flow constructedwetlandFlow cytometryBacteriaViabilityDecay

A B S T R A C T

In this study the functional status of bacterial biomass within a vertical subsurface flow (VSSF)constructed wetland was examined with the aim to understand the relationship between viable and deadbacteria in soil and influent/effluent wastewater and elucidate the large amount of dead cells in the soilwhich may affect the long-term behavior of the system. The quantification of viable and dead bacteria ininfluent and effluent wastewater and in the soil of a VSSF was performed at single-cell level by flowcytometry (FCM). An optimised pre-treatment was applied to soil samples using sodium pyrophosphateand ultrasonication at a specific energy of 80 kJ/L. Viable and dead cells were detected on the basis ofcellular membrane integrity coupling SYBR-Green I and Propidium Iodide. The bacteria profile in the VSSFsoil depends on the depth and the material grain size. In the upper 0–10 cm sand layer the number of totalbacteria per gram of dry weight (DW) was higher (1.82 � 109 cells/gDW) than in the deeper 40–50 cm(4.8 � 108 cells/gDW) probably due to the vertical feeding and a sieving effect of influent in the top layers.Bacterial biomass in the entire VSSF depth was 0.082 mgVSS/gDW or 144 gVSS/m3 (per cubic meter ofVSSF bed). Size of viable bacteria in the VSSF was smaller (0.16 mm3/cell) than typical size of activatedsludge (0.23 mm3/cell), due to lower nutrient conditions and a longer retention time of viable bacteria inthe bed, estimated at around 130 days by mass balance. Dead bacteria were prevalent in the VSSF soilwith a viable/dead bacteria ratio (V/D) of 0.52. The content of dead bacteria might be higher in the soildue to the presence of unsaturated zones not reached by fresh influent wastewater (“dead-zones”),where moisture and substrate are not so available and bacteria may die. Conversely, the higher V/D ratio(3.3) in the effluent reflects the enrichment of wastewater with viable bacteria during the passagethrough the VSSF bed and along preferential water flow, with higher water content and substrateavailability, where the bacterial growth is favored.

ã 2015 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Ecological Engineering

journal homepage: www.else vie r .com/ locat e/e coleng

1. Introduction

The removal of pollutants in Constructed Wetlands (CW) ismostly driven by microorganisms which are closely tied to thecycling of carbon, nitrogen and sulfur (Faulwetter et al., 2009).Organic matter removal is carried out by heterotrophic bacterialiving under both aerobic and anaerobic conditions, while nitrogenremoval in CWs is the result of the combination of nitrification anddenitrification, even though the significant role of anaerobicoxidation of ammonium was demonstrated in the last years (Truuet al., 2009). Although the high relevance of bacteria communitiesthat live in CW and in vertical subsurface flow CWs (VSSF) is widelyrecognised and well documented (Faulwetter et al., 2009; Truu

* Corresponding author. Tel.: +39 0461 282683.E-mail address: paola.foladori@ing.unitn.it (P. Foladori).

http://dx.doi.org/10.1016/j.ecoleng.2015.04.0580925-8574/ã 2015 Elsevier B.V. All rights reserved.

et al., 2009; Kadlec and Wallace, 2009), the quantification and thedistribution of bacteria in such systems (both in water and in soil)have not been investigated sufficiently and some microbiologicalaspects remain unknown (Tietz et al., 2008; Langergraber, 2011).

Conventional cultivation-based methods produce heavilybiased results and strong underestimations when applied toquantifying bacteria in CWs due to the uncultivability of a largefraction of bacteria (Decamp and Warren, 2001; Alexandrino et al.,2007).

Molecular and biochemical approaches can lead to more in-depth microbiological investigations, but they are not yetcompletely exploited in the CWs field (Faulwetter et al., 2009).Among them, fluorescent staining of microbial nucleic acids,fluorescent in-situ hybridisation of 16S rRNA gene sequences(FISH), polymerase chain reaction (PCR or PCR-DGGE) or newemerging metagenomic approach have been proposed in theliterature to investigate microbial diversity and abundance in

Table 1Concentrations of the main parameters in influent and effluent wastewater in theVSSF CW.

Influent concentration(mg/L)

Effluent concentration(mg/L)

Removal rate(g m�2 d�1)

COD 544 89 28.9Filtered COD 282 51 14.7TSS 161 41 7.6TKN 72.7 17.1 3.5NH4-N 59.7 12.9 3.0NO3-N 2.4 32.6 –

Total P 8.6 5.4 0.2No. samples 55–84 55–85 –

50 P. Foladori et al. / Ecological Engineering 82 (2015) 49–56

wastewater treatment and CWs (Decamp and Warren, 2001;Sawaittayothin and Polprasert, 2007; Alexandrino et al., 2007;Tietz et al., 2008; Krasnits et al., 2009; Faulwetter et al., 2009; Zhaoet al., 2010; Zhang et al., 2012; Adrados et al., 2014; Ligi et al., 2014;Lv et al., 2014). In these microbiological investigations, bacteriaviability and death have rarely been quantified (Decamp andWarren, 2001). When the absolute quantification of bacteria isdone under epifluorescence or confocal microscope, the analysesresult as time-consuming and labor-intensive, requiring severalhours to analyse few samples.

Flow cytometry (FCM) has been proposed for the absolutequantification of bacteria in environmental samples at a rate ofthousands of cells in few minutes, with high accuracy andprecision (Porter et al., 1997; Steen, 2000; Vives-Rego et al.,2000; Bergquist et al., 2009; Foladori et al., 2010). FCM is a multi-parametric, single-cell and rapid analysis which acquires, simul-taneously, fluorescent signals (related to cell properties) and lightscattering signals (related to cell size and biovolume) for eachbacteria cell. In the field of CWs, FCM has rarely been exploited toinvestigate bacteria abundance or to distinguish specific propertiesof bacteria such as viability, activity or decay and only rarepublications are available in the field (Scholz et al., 2001; Gagnonet al., 2007; Chazarenc et al., 2009).

The objective of this study was to explore the functional statusand the depth profiles of bacterial biomass within a VSSF CW, andto compare them with the biomass in the influent and effluentwastewater. The ultimate objective is to provide a betterunderstanding of which portion of the bacterial biomass isresponsible for the turnover of elements (living or viable biomass)and which portion is inactive in the soil (dead biomass, whichmight have a role in the long-term clogging mechanisms).

In this study, viability and death of bacteria were investigatedapplying FCM with the aim of evaluating: (1) the time-profiles ofviable and dead bacteria in influent and effluent wastewater duringtypical cycles; (2) the depth-profiles of viable and dead bacteria ingravel and sand layers; (3) some aspects affecting the physiologicalstatus of bacteria cells when removed from wastewater andretained in the soil.

To our knowledge, the FCM analysis of viability and decay ofbacteria in VSSF CW systems, comparing simultaneously influentwastewater, effluent wastewater and bacteria attached to soilinside the bed, has not been reported yet in the literature, whileDecamp and Warren (2001) analysed viable and dead bacteria onlyin influent and effluent wastewater. This paper aims to contributeto advancing knowledge in microbiological aspects of VSSF CWs byconsidering viable and dead bacteria both in water and soil and byproviding a bacteria biomass balance and some aspects about theretention of bacteria in the VSSF system.

2. Materials and methods

2.1. The VSSF CW plant

The outdoor above-ground VSSF pilot plant was located at739 m a.s.l. (Province of Trento, Italy). The screened municipalwastewater underwent settlement in an Imhoff tank before beingpumped up into the VSSF wetland. The VSSF wetland had a surfacearea of 2.25 m2 and was characterised by the following layers,starting from the bottom: (1) 0.2 m coarse gravel Ø15–30 mm, (2)0.05 m medium gravel Ø7–15 mm, (3) 0.5 m sand Ø1–3 mm, (4)0.05 m medium gravel Ø7–15 mm (layers are graphically indicatedin Supplementary material SPM1). Porosity of both sand and gravelwas 0.30.

The VSSF was a conventional down-flow system with a singlefeeding per cycle applied on top of the bed (6.6 h/cycle, 3.6 cycles/day) and wastewater was discharged by free drainage for the entire

cycle, resulting in a time-variable flow rate effluent during thecycle. The hydraulic loading rate applied to the VSSF was 63 L m�2

d�1 on average during the 1-year monitoring period, while theapplied organic load was correspondent to 3.2 m2/PE. The plant hasbeen in operation since 2009 and this research was carried outafter three years of operation. The VSSF CW was unplanted, butTietz et al. (2007) showed no statistically significant difference inbacterial biomass in all the layers of planted and unplanted VSSFsystems.

2.2. Sampling and chemical analyses in influent and effluentwastewater

Samples of influent and effluent wastewater were collectedweekly. Effluent concentrations were measured in 6.6-h-compos-ite-samples obtained by sampling aliquots proportional to effluentflow rate. Intensive monitoring campaigns were conductedmonthly during the typical VSSF cycle to obtain time-profiles ofeffluent concentrations in the intervals 0–5 min, 5–10 min,10–20 min, 20–30 min, 30 min�1 h, 1–2 h, 2–4 h, 4–6 h and 6 h.

Routine analyses were: total Chemical Oxygen Demand (COD),Total Suspended Solids (TSS), Total Kjeldhal Nitrogen (TKN), NH4-N, NO3-N and total P (APHA, 2005). Filtered COD was measuredafter filtration on a 0.45-mm-membrane. The mean concentrationsin influent and effluent wastewater are indicated in Table 1 andthey are in accordance with typical performances expected fromVSSF systems. The VSSF wetland demonstrated high efficiency inTSS removal (>74%), total COD and filtered COD removal (>81%).Nitrification occurred in the VSSF wetland due to the aerobicconditions originated by the intermittent loading mode. Influentwastewater temperature varied in range from 7 to 20 �C during oneyear of monitoring.

2.3. Plate counts

Total Coliforms, Escherichia coli and Faecal Streptococci, whichare common microbiological indicators in wastewater, weremeasured for a comparison in influent and effluent wastewaterby Membrane Filter Techniques. The original samples were dilutedto 10�2 to 10�6 and volumes of 10–100 mL (in triplicate) werefiltered on 0.45-mm membrane filters to obtain plates with20–100CFUs.

Total Coliforms were measured after growth on Coliforms-E.coli agar (C-EC agar, Biolife, Italy) for 24 � 2 h at 36 � 1 �C (ItalianStandards APAT CNR IRSA Manual 29/2003 Methods 7010C). E. coliwere measured after growth on Tryptone Bile X-glucuronidemedium (TBX, Oxoid, UK) after 21 �3 h at 44 �1 �C according toItalian Standards (APAT CNR-IRSA, 2003a). Faecal Streptococciwere measured by growth on Slanetz and Bartley medium (Oxoid,UK) for 44 � 4 h at 36 � 1 �C according to Italian Standards (APATCNR-IRSA, 2003b).

P. Foladori et al. / Ecological Engineering 82 (2015) 49–56 51

2.4. Outline of bacterial soil detachment

Samples of soil (granular medium within the bed) were takenrandomly in the VSSF wetland by a corer. Six sampling campaignswere performed. After sampling, the soil cores were transported tothe lab to perform FCM analysis on the same day. Aliquotes of soilunderwent dry-weight analysis (DW) carried out at 105 �C until aconstant weight was achieved.

2.4.1. Washing cyclesFor the detachment of bacteria from the soil, a 0.1 M sodium

pyrophosphate solution (Na4P2O7 solution) was used to loosen thestrong hydrogen bindings and van der Waals, electrostatic andchemical forces that tie cells and particles together (Amalfitanoand Fazi, 2008; Tietz et al., 2007). Briefly, a 100-mL-flask was filledwith 5–20 g of wet soil (Fig. 1, step A) and then with 50 mL ofNa4P2O7 solution for the first washing step (Fig. 1, step B). The flaskwas placed in a shaker at 150 rpm for 15 min. The entire liquid partwas then separated from the soil, and the soil underwent a secondwashing step with a further amount of 50 mL Na4P2O7 solution for15 min. Finally, the liquid part was completely separated from thesoil and added to the liquid part separated in the first washing step,thus obtaining a final suspension with a volume of 100 mL (Fig. 1,step C).

To confirm the maximum recovery of bacterial cells from soilsamples, fluorescent microscopic analyses were performed on soilsamples after washing steps.

Step 100-mL-flask filled wit

Step Add ition of 50 mL N

and washi

Step Fina l suspen sion (1

Dil ution and fil tration on 20 -µm-fil ter

Step E.Fluo rescen t staining of cell s with

SYBR-Green I + Prop idium Iod ide

Numbe r of viab leand dead ba cteria

Step F.FCM ana lysis ( > 20 ,000 cell s)

Biomass of viab le and dead bacteria (as COD)

Biovolumeof ba cteria

Cores ofVSSF-CW

StepSon ication at E

Influ ent andefflu ent was tewater

Fig. 1. Flow chart of the pre-treatment, FCM analysis and bi

2.4.2. Sample sonicationThe bacterial suspension underwent a dispersion step with the

aim of separating bacteria aggregates and obtaining a single-cellsuspension suitable for FCM analysis. The flask containing the 100-mL-suspension, surrounded by cooling water to avoid an increaseof temperature, underwent sonication (Branson 250 Digital Ultra-sonifier, frequency of 20 kHz) at specific energy (Es = time � power/volume of liquid) of up to 100 kJ/L (Fig. 1, step D). The power usedfor Es calculation was the actual power transferred to the bacterialsuspension measured by calorimetry (Mason et al., 1992). Asshown in Fig. 2, the optimal Es value of 80 kJ/L (Es = 210s � 38W/100 mL) permitted the recovery of the maximum number offree viable bacteria minimising the disruption of bacteria. Thisvalue is in agreement with the optimal Es value found for activatedsludge by Foladori et al. (2007).

The influence of the wet soil weight placed in the 100-mL-flaskon the accuracy of the FCM analysis was evaluated calculating thecoefficient of variation (CV) on three replicates for various amountsof soil (sand in particular) used in the analysis (details are indicatedin Supplementary material SPM3). A weight equal or greater than10 g had to be used to obtain a CV lower than 10% in themeasurement of viable and dead cells.

2.5. Fluorescent staining and FCM analysis

FCM analysis was applied to: (1) influent and effluentwastewater collected as described in Section 2.2; (2) bacteriasuspensions recovered from VSSF soil as described in Section 2.4.

A.h 5-20 g of wet soil

B. a4P2O7 solution

ng (2x)

C.00 mL volume)

the soi l

D.s of 80 kJ /L

Dry-weigh t analysis

Pre-treatm

ent

of th

e soil sam

ples

FC

M an

alysisC

alculatio

n o

f b

acteria bio

mass

omass calculation in VSSF soil samples and wastewater.

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 10 0 120 140

Fre

e vi

able

cel

ls (

% o

f N m

ax)

Spe cifi c energy, Es (kJ L-1 )

activated sludg e

bacteria detached an ddisag gregated f rom theVSSF soi l (this paper)

bacteria disagg rega ted from activated sludge (da ta f rom Folado ri et al., 2007 )

Fig. 2. Free viable bacteria detached and disaggregated from the VSSF soil as afunction of Es and comparison with disaggregation of activated sludge.Free viable cells are expressed as a percentage of the maximum number of cellsrecovered (Nmax).

52 P. Foladori et al. / Ecological Engineering 82 (2015) 49–56

All samples were diluted in Phosphate Buffer Saline (PBS) toobtain a concentration of 103 to 104 cells/mL, suitable for FCManalysis. Aliquots of 1 mL of these suspensions were filteredthrough 20-mm filters with the aim of excluding coarse solids (forexample silt) which risk clogging the cytometer nozzle.

Viable and dead bacteria cells were identified on the basis oftheir cellular membrane integrity, applying the double fluorescentstaining with Propidium Iodide and SYBR-Green I (Ziglio et al.,2002): viable cells are defined as those with an intact membrane,while cells with a permeabilised membrane are classified as deadcells (Nebe-von-Caron et al., 2000). In particular, cells were stainedwith 10 mL of Propidium Iodide (PI, stock solution concentration1 mg/mL; Invitrogen, USA; lex = 536 nm, lem = 617 nm) and 10 mLof SYBR-Green I (SYBR-I, 1:30 dilution of commercial stock;Invitrogen, USA; lex = 495 nm, lem= 525 nm) and incubated in thedark for 15 min at room temperature prior to FCM analysis (Ziglioet al., 2002). SYBR-I is capable of staining all cells, whereas thepolarity of PI allows it to penetrate only cells with permeabilisedmembranes (dead cells). In dead cells, the simultaneous stainingwith SYBR-I and PI activates energy transfer between the dyes and,consequently, viable bacteria emit green fluorescence whilst deadbacteria emit red fluorescence (Fig. 1, Step E).

Stained bacteria were analysed with an Apogee A40 flowcytometer (Apogee Flow Systems, UK) equipped with an Ar laser at488 nm (Fig. 1, Step F). The counting rate was of about 100 cells/sand more than 20,000 cells were analysed for each sample toensure accurate results. Thresholds were set on green fluorescenceand red fluorescence in order to exclude non-fluorescent debris.

Fig.3. FCM dot plot of red versus green fluorescence to distinguish viable and dead bactereffluent wastewater.

Regions were identified in the dot plots to distinguish viable anddead cells (Fig. 3).

Green and red fluorescences were collected with logarithmicgain, while the forward angle light scattering (FALS) was collectedwith linear gain, in order to permit the conversion of FALS intobacteria biovolume (data not shown) according to Foladori et al.(2008).

A Student’s t test, producing a p value, was applied to evaluatethe significant difference in the number of bacteria and thebacteria biovolume, assuming significance at p < 0.05.

2.6. Calculation of bacteria biomass

The biomass of each bacterial cell (M) was expressed asChemical Oxygen Demand (COD) as indicated in the best-knownmathematical models used to describe biochemical transforma-tion and degradation processes in subsurface flow constructedwetlands (Langergraber et al., 2009). M was calculated taking intoaccount the biovolume (V) of each bacteria, according to thefollowing expression:

M½gCOD� ¼ V � Cs � 10�15 � 1:480:53

where the carbon content per unit of cell volume (Cs) is 310 fgCmm�3 (Fry, 1990), the carbon content is 53% of dry weight of cells(derived from the empirical formula of bacteria composition,C5H7NO2) and the coefficient 1.48 is used to convert the dry weightof cells into COD. Finally, the entire bacteria biomass in a samplewas calculated summing up M of all bacteria cells (Foladori et al.,2010) and expressing it per gram of DW (gCOD/gDW) or per unit ofsoil volume (gCOD/m3).

3. Results and discussion

3.1. Time-profiles of viable and dead bacteria in the influent andeffluent wastewater

The profiles of viable and dead bacteria in the effluent from theVSSF wetland after the feeding of influent wastewater, measured indifferent runs are shown in Fig. 4. The highest presence of deadbacteria was observed in the influent wastewater taken from theImhoff tank used as a pre-treatment. After only 5 min of feeding,viable and dead bacteria in the influent wastewater (9.9 � 1010 and7.7 � 1010 cells/L, respectively, on average in all the monitoredruns) were reduced immediately in the effluents discharged fromthe VSSF wetland by gravity, with a reduction always above 40%and 80% for viable and dead bacteria respectively (Fig. 4). At

ia in water and soil: (A) influent wastewater; (B) 0–10 cm sandy layer in the VSSF; (C)

Fig. 4. Profiles of viable and dead bacteria in the VSSF in two different runs (1-year monitoring).

P. Foladori et al. / Ecological Engineering 82 (2015) 49–56 53

10–20 min after feeding the concentration of viable and deadbacteria in the effluents formed a peak. The highest flow ratedischarged from the VSSF bed at this time increased tangentialshears, the detachment of bacteria from the granular medium andan increase of the bacteria concentration in the effluent.Successively, the bacteria concentration in the VSSF effluentsdecreased progressively, both for viable and dead cells (Fig. 4), dueto the lower flow rate of the wastewater percolating through thevertical bed and a limited detachment of bacteria retained in thebed.

During the typical 6.6-h-cycle, the progressive percolationof wastewater through the VSSF bed caused an increase inthe proportion of viable bacteria in the effluent and an increase inthe ratio of viable/dead bacteria (V/D), which was about 1.0 in theinfluent wastewater, but increased remarkably up to 3.0–3.9 in theeffluent in the first 30 min, indicating a higher amount of viablebacteria in the effluent.

The concentration of viable bacteria measured by FCM in theinfluent and effluent wastewater was compared with somegroups of culturable faecal indicators, often considered inroutinary microbiological analyses (Table 2, data of 1-yearmonitoring). In average samples weighted by flow rate, a removalefficiency of 84–95% for total coliforms, E. coli and faecalstreptococci was observed in the VSSF CW, in agreement withother findings in the literature reporting efficient removal of about

Table 2Viable and total bacteria measured by FCM in influent and effluent wastewater (average sand relative ratios.

Influent

FCM Viable bacteria (cells/L) 9.3 � 101

Total (viable + dead) bacteria (cells/L) 1.7 � 101

Plate counts Total coliforms (CFU/L) 5.2 � 108

Escherichia coli (CFU/L) 1.1 �108

Faecal streptococci (CFU/L) 4.2 � 107

Ratios Total coliforms/Viable bacteria x 100 0.56%

Escherichia coli/Viable bacteria x 100 0.12%

Faecal streptococci/Viable bacteria x 100 0.05%

0.8–1.7 log-units (Arias et al., 2003). The concentrations of viablebacteria and the culturable groups differed demonstrably by 2–3orders of magnitude in the influent and of 3–4 orders in theeffluent. The percentage of culturable faecal indicators withrespect to viable bacteria was higher in influent wastewater, whileit was lower in the effluent due to two main phenomena: (i)retention of a part of influent bacteria, including culturable faecalindicators, in the bed and enrichment of effluent with viablebacteria grown directly in the bed itself; (ii) loss of cultivability of apart of influent bacteria, especially those of faecal origin, whichmay enter a “viable-but-not-culturable” state (VBNC state) beforebeing released into the effluent wastewater. The significantprecence of VBNC bacteria in CW effluent was highlighted severaltimes (inter alia Decamp and Warren, 2001; Alexandrino et al.,2007). The release in the effluent of viable bacteria of non-faecalorigin is the reason for their apparently lower removal efficiencythan the cultivable ones; similar observations were made byDecamp and Warren (2001).

Bacteria size and biovolume of bacteria in influent and effluentwastewater was estimated from FALS signals acquired by FCMaccording to Foladori et al. (2008). The mean biovolume of viablebacteria in the influent wastewater was 0.19 mm3/cell (diameter ofthe equivalent sphere was 0.71 mm), significantly higher than thebiovolume of viable bacteria in the VSSF effluent (0.17 mm3/cell)with similar values during the entire 6.6-h-cycle.

amples weighted by flow rate), compared with groups of culturable faecal indicators

concentration Effluent concentration Removal efficiency0 3.6 � 1010 62%1 4.5 �1010 73%

2.6 � 107 95%9.2 � 106 92%6.8 � 106 84%0.07% –

0.02% –

0.02% –

54 P. Foladori et al. / Ecological Engineering 82 (2015) 49–56

3.2. Depth-profiles of viable and dead bacteria in the VSSF granularmedium

The profiles of viable and dead bacteria at various depths of theVSSF CW are shown in Fig. 5. The depth has a great influence on thedistribution of bacteria content and this is consistent with otherobservations in the literature (inter alia Tietz et al., 2008).

The bacteria content in the upper 5 cm gravel layer (Ø7–15 mm)was significantly lower compared to the 0–10 cm sand layer (Ø1–3 mm), which offers a larger surface per unit of dry weight forbacterial attachment. The highest content of viable bacteria wasfound in the upper 0–10 cm sand layer, as a consequence of twomain phenomena: (i) the vertical feeding which causes a higheravailability of organic matter and nutrients and a consequentstimulation of microbial growth (Tietz et al., 2008; Faulwetteret al., 2009); (ii) the filtration of influent wastewater and theentrapment of bacteria embedded in solids.

The V/D ratio was calculated as viable cells over dead cells. Inthe upper 5 cm gravel layer the ratio V/D was low (0.38), due to thefact that a larger fraction of bacteria entering the wetland with thewastewater are already dead. Truu et al. (2009) indicated that deador damaged bacteria are removed from the wastewater moreefficiently than undamaged cells.

The V/D ratio increased in the upper 0–10 cm sand layer, whereV/D ratio was 0.73, which was the maximum value along the depthof the bed. In this layer, the growth of viable bacteria is favored dueto the higher availability of organic matter and nutrients and thelarger surface for bacterial attachment. In the layers from 10 to50 cm the V/D ratio was always lower than 0.5. The mean V/D ratioin the entire bed was 0.52. The considerable accumulation of deadbacteria in the VSSF soil, as shown in Fig. 5, suggests that bacteriaundergo slow cellular lysis or are hard to lyse and take longer todecompose. By comparison, bacteria in activated sludge present aV/D ratio of around 4.1 (Foladori et al., 2010), which is 8 timeshigher the V/D ratio in the VSSF wetland. A slower autolysis rate isprobably the reason of the accumulation of dead cells inside thebed. This may contribute to the mechanisms involved in clogging,in addition to biomass production, deposition/filtration of incom-ing particulates and chemical precipitation (Kadlec and Wallace,2009).

The prevalence of dead bacteria inside the VSSF bed comparedto viable ones was, to our knowledge, not observed before.Microbiological methods applied to CWs are in fact not designedspecifically to detect dead cells, but are used mainly to quantifyvital cellular properties (such as the presence of RNA or ATP

Fig. 5. Profiles of viable and dead bacteria at various depths of the VSSF wetland(upper gravel layer and main sand layer are distinguished) expressed per unit of dryweight (DW) of the soil.

content) or to count total cells independently on their viability ordeath (such as DNA staining or some PCR-based methods).However, Sawaittayothin and Polprasert (2007), using FISHanalysis of bacteria populations in CWs, indicated that the EUB338 mix probes (specific for 16S rRNA) detected about 50% of totalbacteria identified with DAPI staining, probably suggesting that theremaining 50% could be made up of inactivated or dead bacterialcells. In a horizontal subsurface CW with relatively high organicloads, Krasnits et al. (2009) reported that viable cells (FISH-stainedcells) represent 60–80% of the total cells enumerated by DAPIstaining.

In the sand layer the mean biovolume of the dead bacteria was0.14 mm3/cell, whilst that of the viable bacteria was 0.16 mm3/cell.These values are similar to the biovolume of 0.13 mm3/cellindicated by Tietz et al. (2007) in wetlands for bacteria observedunder epifluorescence microscope. By comparison, bacteria inactivated sludge have a biovolume of 0.23 mm3/cell (Foladori et al.,2010) or 0.25 mm3/cell (Frølund et al., 1996), which is consistentlylarger than the biovolume of bacteria in the VSSF wetlands. Thesmaller size may be due to the development of bacterialcommunities with different cell size and to an environment inCWs with relatively lower nutrient content than activated sludge.As an example, the volumetric organic load applied in this VSSFwetland was approximately 0.1 kgCOD m�3 d�1, while activatedsludge for carbon oxidation and nitrification is exposed to typicalloads of around 1.0 kgCOD m�3 d�1. In general, it is well known thatbacteria reduce their size as a consequence of nutrient limitedenvironments (Robertson et al., 1998; Blagodatskaya and Kuzya-kov, 2013).

The amount of total bacteria in the VSSF soil was calculated asthe sum of both viable and dead cells. The density of total bacteriaat various depths of the VSSF system was compared to the resultsobtained by Tietz et al. (2008) in unplanted VSSF wetlands (thecomparison is graphically shown in Supplementary materialSPM4). Both total bacterial profiles appear similar except for thefirst sand layer, which could be attributed to the different grain sizeof sand used in these two researches: sand Ø1–3 mm and d10 of1.0 mm in this work and sand Ø0.06–4 mm and d10 of 0.2 mm usedby Tietz et al. (2008). The use of sand with a smaller d10 mayenhance the filtration and the accumulation of influent bacteriaand organic solids in the upper centimeters of the bed. Conversely,in a coarse sand, the transport and penetration of bacteria throughthe bed is increased. A potential advantage of the use of coarsematerial is the reduction of clogging risk in the upper layers.However, a disadvantage may be a higher effluent concentration ofbacteria due to the reduced retainment capacity of the bed. Themean density of total bacteria in the entire depth (weighted by thelayer depths) of our VSSF system was 1.0 � 109 cells/gDW, veryclose to the value obtained by Tietz et al. (2008) (1.4 �109 cells/gDW).

3.3. Comparison between viable and dead bacteria in wastewater andin soil

The bacteria biomass in wastewater and soil, quantifiedaccording to the Section 2.6, was expressed per unit of DW orper unit of volume in the VSSF. The size of bacteria and the biomassestimations have rarely been reported in the CW field (Tietz et al.,2007; Tietz et al., 2008).

The mean biomass of total bacteria (viable + dead) in the VSSFsoil was 213 gCOD/m3 and 0.12 mgCOD/gDW (weighted by thelayer depths). This latter value is very similar to the value ofapproximately 0.10 mgCOD/gDW calculated by us from the data ofTietz et al. (2007). These results are in the range of biomass valuesin VSSF systems reviewed by Truu et al. (2009) of 0.022–1.69 mgC/gDW, corresponding to about 0.06–4.6 mgCOD/gDW. For instance,

P. Foladori et al. / Ecological Engineering 82 (2015) 49–56 55

the total amount of microbial biomass expected in common soilsranges from 0.1 to 4.0 mg of cell dry weigh per g of soil(Blagodatskaya and Kuzyakov, 2013).

On average, the total bacterial biomass fed into the VSSF CWwith the influent wastewater converted into COD, was 3.1 gCODm�3 d�1 whilst, in the effluent wastewater during the entire 6-hcycle, it was 0.77 gCOD m�3 d�1. The difference between input andoutput (2.3 gCOD m�3 d�1) is the mean total bacteria biomassremoved every day from wastewater and accumulated in the VSSFbed. This corresponds to 1.1% of the total biomass present in thebed which is also produced by the bacteria growth. The totalbacteria biomass daily leaving the VSSF bed with the effluentwastewater was only 0.36% of the total biomass present in the bed.Considering viable bacteria, the viable biomass removed every dayfrom wastewater was 1.5% of the viable biomass in the bed, whilethe viable biomass left in the effluent was 0.77%. Although theconcept of solid retention time (sludge retention time or sludgeage, common in activated sludge systems; Henze et al., 2008) is notused in wetlands, a ballpark balance of viable biomass was herecalculated, obtaining a rough mean retention time of the viablebiomass in the bed of around 130 days. For instance, this value is farlonger than solid retention time in activated sludge systems, wheresludge age is typically in the range of 5–25 days. It has beendemonstrated that the observed growth yield in biologicalprocesses is inversely dependent on the solid retention time andendogenous respiration (Lawrence and McCarty, 1970; vanLoosdrecht and Henze, 1999; Liu and Tay, 2001). A decrease inthe observed growth yield of the bacterial biomass is thus expectedin the VSSF bed as a consequence of the longer solid retention timein these systems.

3.4. Hypothesis of dead and viable bacterial distribution inunsaturated VSSF soil

The mean V/D ratio (viable/dead cells) in the soil was 0.52,much lower than the V/D ratio in the influent wastewater (0.93 onaverage) and in the effluent wastewater (3.3 on average). This largediscrepancy indicates that the dead cells accumulated within theVSSF soil, while effluent wastewater was remarkably enriched inviable cells.

These observations could be explained assuming the hypothesisof preferential flow paths formed during percolation through theunsaturated VSSF soil (a theoretical schema of this hypothesis isshown in Supplementary material SPM5). The wastewater flowpath may follow preferential directions due to the variablemoisture content, the presence of small channels closed to flowand the variable biofilm thickness around the sand grains. Theconsequence of this is the possible formation of: (1) zones ofpreferential water flow, with high water content and substrateavailability, where the growth of viable bacteria is favored; (2)zones where fluid and air arrive randomly or not at all and thesubstrates are limited or absent, where bacteria can not grow,decay and die.

The presence of “dead zones” in VSSF systems was observed byGiraldi et al. (2009) as a result of hydrodynamic studies whereactual and theoretical residence times were compared. Deadzones, caused by preferential water flow, biofilm accumulation,porosity decrease or clogged regions, were predicted by Rajabza-deh et al. (2015) in a VSSF system using a calibrated model. It isbelieved that organic matter flux, required for the growth ofheterotrophic bacteria, may be limited in dead zones, where alsooxygen could become a limiting component (Rajabzadeh et al.,2015); the result may be a limitation in bacteria survival causing asubsequent death. Giraldi et al. (2009) identified dead zones in theupper layers where water flows near the distribution holes(individual points), while Rajabzadeh et al. (2015) identified dead

zones near the bottom in an area opposite the outlet where thewater velocity is low. Understanding and preventing conditionswhich lead to dead zones and channeled flow in CWs, for instanceusing multiphysics modelling, permits to avoid negative effects onthe treatment efficiency of the system as a whole (Rajabzadehet al., 2015).

Conversely, bacteria growth seems rather favored in water filmsalong the preferential water flow directions. Here the fluid velocitycauses tangential shear and detachment of bacteria from the zoneswhere the viable bacteria are more abundant. The consequence ofthis elution of viable bacteria from the soil is the shift in theeffluent microbial population which will be enriched in viablecells, whilst dead bacteria remain confined in the dead zonesexcluded from the main water flow. The V/D ratio in the effluentwill subsequently appear higher than the V/D ratio in the soil andalso than the V/D in the influent wastewater.

The hypothesis that bacteria in the effluent wastewater maylargely originate from the detachment of biofilm grown on soilrather than as a fraction of the influent bacterial community wasconfirmed recently by Adrados et al. (2014). Adrados et al. (2014),using the PCR-DGGE molecular technique, observed that in VSSFwetlands there was almost no relation between the influent andthe effluent bacterial communities, suggesting that the communitystructure in the system and in the effluent wastewater is related toother conditions developed inside the bed rather than the influentwastewater microbiology.

4. Conclusions

In this study, viable and dead bacteria were quantified by CM ininfluent, effluent wastewater and in the VSSF soil. The mainconclusions are:

1) The study has confirmed that bacteria profile in the VSSFsystems depends on the depth: the number of viable cells in theupper 0–10 cm sand layer was 3.7 times higher than in thedeeper 40–50 cm, due to the vertical feeding and a sieving effectof influent in the top layers.

2) The study has confirmed the removal in the VSSF system of asignificant part of influent bacteria (viable bacteria andculturable faecal indicators);

3) The long retention time of the viable biomass in the VSSF soil(130 days) explains the low observed growth yield of bacterialbiomass is such systems.

4) Dead bacteria accumulate in the VSSF soil: the number of deadcells in the soil were 2 times the viable ones, due to slowautolysis rate and with a possible influence on the long-termbehavior of the VSSF systems.

5) In the effluent the viable bacteria were 3.3 times the dead ones:effluent wastewater transports viable bacteria during thepassage through the VSSF bed, probably along preferentialwater flow where the bacterial growth is favored. Thehypothesis that dead bacteria may remain confined in the deadzones excluded from the main water flow was proposed.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.ecoleng.2015.04.058.

References

Adrados, B., Sánchez, O., Arias, C., Becares, E., Garrido, L., Mas, J., Brix, H., Morato, J.,2014. Microbial communities from different types of natural wastewater

56 P. Foladori et al. / Ecological Engineering 82 (2015) 49–56

treatment systems: vertical and horizontal flow constructed wetlands andbiofilters. Water Res. doi:http://dx.doi.org/10.1016/j.watres.2014.02.011.

APAT CNR-IRSA, 2003a. Methods 7030 F. In Metodi analitici per le acque. Manuale29/2003, Sezione 7000, Metodi per microrganismi. In Italian.

APAT CNR-IRSA, 2003b. Methods 7040C. In Metodi analitici per le acque. Manuale29/2003, Sezione 7000, Metodi per microrganismi. In Italian.

APHA, AWWA, and WPCF, 2005. Standard Methods for the Examination of Waterand Wastewater. 21st edition. Washington DC, USA.

Alexandrino, M., Grohmann, E., Szewzyk, R., Szewzyk, U., 2007. Application ofculture-independent methods to assess the bacteria removal efficiency ofsubsurface flow constructed wetlands. Water Sci. Technol. 56 (3),217–222.

Amalfitano, S., Fazi, S., 2008. Recovery and quantification of bacterial cellsassociated with streambed sediments. J. Microbiol. Methods 75, 237–243.

Arias, C.A., Cabello, A., Brix, H., Johansen, N.-H., 2003. Removal of indicator bacteriafrom municipal wastewater in an experimental two-stage vertical flowconstructed wetland system. Water Sci. Technol. 48 (5), 35–41.

Bergquist, P.L., Hardiman, E.M., Ferrari, B.C., Winsley, T., 2009. Applications of flowcytometry in environmental microbiology and biotechnology. Extremophiles 13(3), 389–401.

Blagodatskaya, E., Kuzyakov, Y., 2013. Active microorganisms in soil: critical reviewof estimation criteria and approaches. Soil Biol. Biochem. 67, 192–211.

Chazarenc, F., Gagnon, V., Comeau, Y., Brisson, J., 2009. Effect of plant and artificialaeration on solids accumulation and biological activities in constructedwetlands. Ecol. Eng. 35 (6), 1005–1010.

Decamp, O., Warren, A., 2001. Abundance, biomass and viability of bacteria inwastewaters: impact of treatment in horizontal subsurface flow constructedwetlands. Water Res. 35 (14), 3496–3501.

Faulwetter, J.L., Gagnon, V., Sundberg, C., Chazarenc, F., Burr, M.D., Brisson, J.,Camper, A.K., Stein, O.R., 2009. Microbial processes influencing performance oftreatment wetlands: a review. Ecol. Eng. 35, 987–1004.

Foladori, P., Bruni, L., Andreottola, G., Ziglio, G., 2007. Effects of sonication onbacteria viability in wastewater treatment plants evaluated by flow cytometry –fecal indicators, wastewater and activated sludge. Water Res. 41, 235–243.

Foladori, P., Quaranta, A., Ziglio, G., 2008. Use of silica microspheres havingrefractive index similar to bacteria for conversion of flow cytometric forwardlight scatter into biovolume. Water Res. 42 (14), 3757–3766.

Foladori, P., Bruni, L., Tamburini, S., Ziglio, G., 2010. Direct quantification of bacterialbiomass in influent: effluent and activated sludge of wastewater treatmentplants by using flow cytometry. Water Res. 44 (13), 3807–3818.

Frølund, B., Palmgren, R., Keiding, K., Nielsen, P.H., 1996. Extraction of extracellularpolymers from activated sludge using a cation exchange resin. Water Res. 30,1749–1758.

Fry, J.C., 1990. Direct methods and biomass estimation. Methods Microbiol. 22,41–85.

Gagnon, V., Chazarenc, F., Comeau, Y., Brisson, J., 2007. Influence of macrophytespecies on microbial density and activity in constructed wetlands. Water Sci.Technol. 56 (3), 249–254.

Giraldi, D., de'Michieli Vitturi, M., Zaramella, M., Marion, A., Iannelli, R., 2009.Hydrodynamics of vertical subsurface flow constructed wetlands: tracer testswith rhodamine WT and numerical modelling. Ecol. Eng. 35, 265–273.

Henze M., van Loosdrecht M.C.M., Ekama G.A., Brdjanovic D. 2008. BiologicalWastewater Treatment: Principles, Modelling and Design. ISBN:9781843391883.

Kadlec, R.H., Wallace, S.D., 2009. Treatment Wetlands, 2nd edition CRC Press ISBN978-1-56670-526-4..

Krasnits, E., Friedler, E., Sabbah, I., Beliavski, M., Tarre, S., Green, M., 2009. Spatialdistribution of major microbial groups in a well established constructedwetland treating municipal wastewater. Ecol. Eng. 35, 1085–1089.

Langergraber, G., Rousseau, D.P.L., García, J., Mena, J., 2009. CWM1: a general modelto describe biokinetic processes in subsurface flow constructed wetlands. WaterSci. Technol. 59 (9), 1687–1697.

Langergraber, G., 2011. Numerical modelling: a tool for better constructed wetlanddesign? Water Sci. Technol. 64 (1), 14–21.

Lawrence, A.W., McCarty, P.L., 1970. Unified basis for biological treatment designand operation. J. Sanitary Eng. Div. 96 (3), 757–778.

Ligi, T., Oopkaup, K., Truu, M., Preem, J.-K., Nõlvak, H., Mitsch, W.J., Mander, U., Truu,J., 2014. Characterization of bacterial communities in soil and sediment of acreated riverine wetland complex using high-throughput 16S rRNA ampliconsequencing. Ecol. Eng. 72, 56–66.

Liu, Y., Tay, J.-H., 2001. Strategy for minimization of excess sludge production fromthe activated sludge process. Biotechnol. Adv. 19, 97–107.

Lv, X., Yu, J., Fu, Y., Ma, B., Qu, F., Ning, K., Wu, H., 2014. A meta-analysis of thebacterial and archaeal diversity observed in wetland soils. Sci. World J. 1–12.doi:http://dx.doi.org/10.1155/2014/437684.

Mason, T.J., Lorimer, J.P., Bates, D.M.,1992. Quantifying sonochemistry: casting somelight on a ‘black art’. Ultrasonics 30 (1), 40–42.

Nebe-von-Caron, G., Stephens, P.J., Hewitt, C.J., Powell, J.R., Badley, R.A., 2000.Analysis of bacterial function by multi-colour fluorescence flow cytometry andsingle cell sorting. J. Microbiol. Meth. 42, 97–114.

Porter, J., Deere, D., Hardman, M., Edwards, C., Pickup, R., 1997. Go with the flow–useof flow cytometry in environmental microbiology. FEMS Microbiol. Ecol. 24, 93–101.

Rajabzadeh, A.R., Legge, R.L., Webe, K.P., 2015. Multiphysics modelling of flowdynamics, biofilm development and wastewater treatment in a subsurfacevertical flow constructed wetland mesocosm. Ecol. Eng. 74, 107–116.

Robertson, B.R., Button, D.K., Koch, A.L., 1998. Determination of the biomasses ofsmall bacteria at low concentrations in a mixture of species with forward lightscatter measurements by flow cytometry. Appl. Environ. Microbiol. 64, 3900–3909.

Sawaittayothin, V., Polprasert, C., 2007. Nitrogen mass balance and microbialanalysis of constructed wetlands treating municipal landfill leachate. Bioresour.Technol. 98, 565–570.

Scholz, M., Xu, J., Dodson, H.I., 2001. Comparison of filter media, plant communitiesandd microbiology within constructed wetlands treating wastewatercontaining heavy metals. J. Chem. Technol. Biotechnol. 76, 827–835.

Steen, H.B., 2000. Flow cytometry of bacteria: glimpses from the past with a view tothe future. J. Microbiol. Methods 42, 65–74.

Tietz, A., Kirschner, A., Langergraber, G., Sleytr, K., Haberl, R., 2007. Characterisationof microbial biocoenosis in vertical subsurface flow constructed wetlands. Sci.Total Environ. 380, 163–172.

Tietz, A., Langergraber, G., Watzinger, A., Haberl, R., Kirschner, A.K.T., 2008. Bacterialcarbon utilization in vertical subsurface flow constructed wetlands. Water Res.42, 1622–1634.

Truu, M., Juhanson, J., Truu, J., 2009. Microbial biomass, activity and communitycomposition in constructed wetlands. Sci. Total Environ. 407, 3958–3971.

van Loosdrecht, M.C.M., Henze, M., 1999. Maintenance, endogenous respiration,lysis, decay and predation. Water Sci. Technol. 39 (1), 107–117.

Vives-Rego, J., Lebaron, P., Nebe-von Caron, G., 2000. Current and future applicationsof flow cytometry in aquatic microbiology. FEMS Microbiol. Rev. 24, 429–448.

Zhang, T., Shao, M.-F., Ye, L., 2012. 454 Pyrosequencing reveals bacterial diversity ofactivated sludge from 14 sewage treatment plants. ISME J. 6 (6), 1137–1147.

Zhao, Y., Liu, B., Zhang, W., Hu, C., An, S., 2010. Effects of plant and influent C:N:Pratio on microbial diversity in pilot-scale constructed wetlands. Ecol. Eng. 36,441–449.

Ziglio, G., Andreottola, G., Barbesti, S., Boschetti, G., Bruni, L., Foladori, P., Villa, R.,2002. Assessment of activated sludge viability with flow cytometry. Water Res.36 (2), 460–468.