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

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

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

Use of silica microspheres having refractive indexsimilar to bacteria for conversion of flow cytometricforward light scatter into biovolume

Paola Foladoria,*, Alberto Quarantab, Giuliano Ziglioa

aDepartment of Civil and Environmental Engineering, University of Trento, via Mesiano, 77, 38050 Trento, ItalybDepartment of Materials Engineering and Industrial Technology, University of Trento, via Mesiano, 77, 38050 Trento, Italy

a r t i c l e i n f o

Article history:

Received 3 January 2008

Received in revised form 5 May 2008

Accepted 23 June 2008

Available online 4 July 2008

Keywords:

Bacteria

Biovolume

Flow cytometry

Forward angle light scattering

Rayleigh–Gans theory

Wastewater

* Corresponding author. Tel.: þ39 461 882 66E-mail addresses: paola.foladori@ing.unit

0043-1354/$ – see front matter ª 2008 Elsevidoi:10.1016/j.watres.2008.06.026

a b s t r a c t

This research describes an alternative approach for the rapid conversion of flow cytometric

Forward Angle Light Scattering (FALS) into bacterial biovolume. The Rayleigh–Gans theory

was considered for explaining the main parameters affecting FALS intensity: sensitivity

analysis of the model was carried out, taking into account the parameters characteristic

of bacterial cells and characteristics of the flow cytometer. For particles with size in the

typical range of bacteria, the FALS intensity is affected mainly by volume and refractive

index of bacterial cells and is approximately independent of the shape of the cells. The pro-

posed conversion from FALS intensity into bacterial biovolume is based on a calibration

curve determined by using silica microspheres having relative refractive index as far as

possible similar to that of bacteria. The approach was validated for two different flow

cytometers (the first equipped with an arc lamp and the second with a laser) by comparing

the biovolume distribution obtained from FALS conversion with the biovolume measured

conventionally under epifluorescence microscopy. The specific case of bacteria taken

from a WWTP was addressed. Compared to the time-consuming conventional microscopic

approach, the application of FALS for sizing bacterial biovolume could be a very promising

tool being completed in few minutes, simultaneously to the enumeration of bacteria during

the flow cytometric analysis.

ª 2008 Elsevier Ltd. All rights reserved.

1. Introduction offers a great potential, thanks to the possibility of measuring

In biotechnological processes and in environmental samples,

viable bacterial biomass is an important issue involved in

mass balance and in evaluating bacteria dynamics (growth

or decay). Cell concentration, biovolume and the specific car-

bon content or the dry weight of cells need to be known for

the calculation of bacterial biomass (Fry, 1990). In environ-

mental samples, cell sizing with epifluorescence or confocal

microscope is the conventional approach for biovolume

determination. Alternatively, the flow cytometry technique

9; fax: þ39 461 882 672.n.it (P. Foladori), quarantaer Ltd. All rights reserved

simultaneously viable bacterial concentration and their for-

ward angle light scattering which is correlated to the cellular

size (or biovolume). Flow cytometry is a single-cell analysis

and is showing increasing interest in environmental microbi-

ology for its rapidity in quantifying microorganisms. This

technique takes only a few minutes for the analysis of sev-

eral hundred or thousand cells in bacterial suspensions

with accuracy and high precision in the enumeration (Porter

et al., 1997; Steen, 2000). In the case of environmental samples,

flow cytometric analysis requires firstly cell staining with

@ing.unitn.it (A. Quaranta), giuliano.ziglio@ing.unitn.it (G. Ziglio)..

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 7 5 7 – 3 7 6 63758

fluorescent probes in order to discriminate microorganisms

from other non-fluorescent non-biotic particles. For example,

the staining of cellular components by fluorescent dyes allows

the total bacteria number to be identified and their viability,

death or metabolic activity to be discriminated. Then total,

viable or dead bacteria can be rapidly and automatically enu-

merated by flow cytometry on the basis of their fluorescence

emission (Nebe-von-Caron et al., 2000; Ziglio et al., 2002). A

flow cytometer is usually equipped for the simultaneous

acquisition of two or more fluorescent signals and two light

scattering signals for each cell passing through the focus. In

particular, the light scattering acquired by the flow cytometer

in the forward direction – that is for angles of about 1–2� and

indicated as FALS (Forward Angle Light Scattering) – depends

on bacteria size (Salzman et al., 1990).

Empirical equations have often been sought by several

authors for the conversion of FALS signal into bacteria size

(Bouvier et al., 2001; Julia et al., 2000; Davey and Kell, 1996).

For microorganisms, most authors agree that FALS increases

monotonically with cell size and is a non-linear function of

the volume or the diameter of the bacteria (Davey and Kell,

1996; Julia et al., 2000). In some cases the non-linear curve

was fitted using a second-order polynomial curve (Davey

and Kell, 1996; Julia et al., 2000).

A rational approach for the conversion of FALS into

biovolume based on the Rayleigh–Gans theory was proposed

by Koch et al. (1996). On the basis of this theory, the FALS in-

tensity is proportional to the sixth power of the equivalent

sphere radius (or to the second power of particle volume).

Several authors mention the possibility of evaluating the

bacteria size by comparing FALS signal produced by the cells

to that given by microspheres (beads) of known diameter

(inter alia Koch et al., 1996). In this procedure, as underlined

by Koch et al. (1996), the refractive index of the series of par-

ticles must be the same. Synthetic beads for flow cytometry

applications are usually made of polystyrene or latex. FALS

signal produced by bacteria generally has a lower intensity

than the one produced by synthetic beads of the same size

or volume as the bacterium, due to the much higher refrac-

tive index of polystyrene or latex than cells. This can induce

an underestimation of the actual biovolume of bacteria, and

therefore the use of latex particles as size standards for bio-

logical cells may be problematic (Davey et al., 1993; Robertson

et al., 1998).

This aspect focuses on the importance of comparing FALS of

particles having the same optical characteristics, especially for

the relative refractive index. FALS intensity produced by cells

in starved strains or in natural communities can differ from

exponentially growing cells, as a consequence of the different

refractive indexes due to the different metabolic status of the

cells grown with excess of substrate or under limiting condi-

tions (Bouvier et al., 2001). Refractive index of bacteria may

vary from 1.36–1.40 for bacterial cells growing in minimal me-

dium or in environments with limited substrate (Valkenburg

and Woldringh, 1984; Robertson et al., 1998) to 1.40–1.41 for

bacilli in cultures (Ross, 1957). Considering marine bacteria

and phytoplankton cells growing in natural environments

where nutrients are less abundant, refractive indices have

been estimated to be in the range 1.39–1.40 (Jonasz et al.,

1997; Morel and Ahn, 1990) and in the range 1.39–1.45,

respectively (Twardowski et al., 2001; Stramski et al.,

2001). Refractive index of marine microorganisms such as

Synechococcus and eukaryotes is 1.41–1.43 on average (Green

et al., 2003). The purified organic material from a typical

phytoplankton cell has an average refractive index of

about 1.53 (Twardowski et al., 2001), so that it is not the

organic material but the large proportion of water that

gives bacteria low refractive indices. We are not aware of

any directly measured value for the refractive index of

bacteria present in wastewater and activated sludge in

the literature. It was estimated in this research on the ba-

sis of the agreement between the FALS converted into bio-

volume and the microscopic sizing of bacteria.

The present research describes a new and alternative

approach to rapidly convert FALS intensity measured by

flow cytometry into bacterial biovolume by using silica mi-

crospheres, having optical characteristics – especially the

relative refractive index – as far as possible similar to those

of bacteria.

The Rayleigh–Gans theory was considered for understand-

ing and supporting the conversion from FALS intensity into

bacterial biovolume, according to the model of Koch et al.

(1996). The Rayleigh–Gans law is applicable only if the phase

shift between the waves scattered from different points of

the target is low. This condition is satisfied if: (a) the size of

bacteria is not too large with respect to the incident light

wavelength and (b) the refractive index of bacteria is similar

to that of the surrounding medium: an index of refraction

only 3–6% higher than that of the surrounding medium mini-

mizes phase change and allows the use of Rayleigh–Gans

theory (Robertson et al., 1998; Koch et al., 1996). Both these

hypotheses are usually satisfied for bacteria in environmental

samples and in WWTPs. A sensitivity analysis was applied to

evaluate how the flow cytometer configuration and the phys-

ical properties of bacteria, both involved in the Rayleigh–Gans

model, affect FALS intensity.

The specific case of bacteria taken from a WWTP was

addressed. The approach was applied to bacteria in wastewa-

ter and activated sludge and was validated by comparing the

biovolume distribution obtained from FALS conversion with

the biovolume measured using a conventional approach

based on microscopy.

By means of the proposed calibration procedure, a rapid

conversion of the FALS signal (measured usually in arbi-

trary units) into biovolume (measured in mm3) is possible,

independent of the geometry and set-up of the specific

cytometer.

2. Materials and methods

2.1. Synthetic beads

Non-fluorescent silica microspheres (produced by MicroParti-

cles GmbH, Germany) with different diameters were used to

assess the calibration curve of FALS intensity. The selected

silica beads have a refractive index of 1.42, which is close to

the typical values of environmental bacteria. Six types of

microspheres with diameter between 0.5 and 1 mm were

chosen. Also fluorescent polystyrene beads (produced by

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 7 5 7 – 3 7 6 6 3759

Spherotech Inc, USA) having diameter ranging from 0.49 to

2.27 mm were used.

2.2. Samples of wastewater and activated sludge

Samples of wastewater were taken after the primary settler of

a municipal wastewater treatment plant (Trento, Italy). Acti-

vated sludge samples were collected from the oxidation tank

of the same plant, operating with a sludge age of about 12

days and at a specific organic load of 0.15 kg BOD5 kg TSS�1 d�1

(BOD5¼ Biochemical Oxygen Demand at 5 days; TSS¼ Total

Suspended Solids). Grab samples were collected and pro-

cessed in the lab within 1–2 h. Samples of wastewater and

activated sludge were diluted 1:20 and 1:500 v/v, respectively,

in PBS (3 g K2HPO4, 1 g KH2PO4 and 8.5 g NaCl L�1; pH¼ 7.2)

reaching approximately a concentration of 106–107 bacteria

per mL. Before the flow cytometric analysis, activated sludge

samples were pretreated in order to disaggregate flocs and

to obtain a suspension of dispersed bacteria. Pretreatment

was performed using sonication at a transferred specific

energy of 80 kJ L�1. This value of specific energy allows the

disaggregation of bacterial flocs avoiding the death of cells

(Foladori et al., 2007).

2.3. Fluorescent staining of cells

Bacteria in wastewater and activated sludge were stained with

the fluorochrome SYBR-Green I (Molecular Probes, Invitrogen),

adding 10 mL of dye (after 1:30 dilution of the commercial stock

solution) to 1 mL of sample containing about 106–107 bacteria.

An incubation of 15 min at room temperature and in the dark

was ensured for the optimal staining of bacteria by the dye.

SYBR-Green I, labelling nucleic acids inside the cells, has

excitation at a wavelength of 488 nm and emission of green

fluorescence at 525 nm (Ziglio et al., 2002).

2.4. Automated cell sizing under microscope

The size distribution of bacterial cells (previously stained with

SYBR-Green I) was measured under microscope, by using

a Nikon Labophot at 1000� magnification and equipped with

epifluorescence apparatus. For the microscopic examination,

a number of 1500–4000 bacteria in about 50–100 fields on

each slide were observed and recorded using a CCD camera

(Photometrics). An image processing program (LUCIA, Nikon)

extracted the particles from the image background after

green-level thresholding (cells stained with SYBR-Green I

emitted green fluorescence) and binarization. For each bacte-

rial cell the main parameters, such as minimum axis (a) and

maximum axis (b) were automatically measured, after the cal-

ibration of the images with a microscopic micrometer. Finally

the bacterial biovolume (V) was calculated, considering the

bacterial shape equivalent to an ellipsoid of revolution (Davey

et al., 1993):

V ¼ 16

pba2 (1)

The measuring of the cell dimensions was done with an error

in minimum and maximum axes of about 4.4% that corre-

sponds to an error in volume of about 13.2%.

2.5. Flow cytometry

Flow cytometric analyses were performed with two different

instruments: (1) a Bryte-HS flow cytometer (Bio-Rad, Inc.,

Hercules, CA) equipped with an arc lamp and is shown in

Fig. 1; (2) an Apogee-A40 flow cytometer (Apogee Flow Sys-

tems, UK) equipped with an Ar laser (488 nm) and character-

ised by an optical configuration similar to the one indicated

in Fig. 1. Apogee’s A40 flow cytometer is a modern high per-

formance model for counting small particles, such as small

bacteria or large virus, thanks to its greater light scatter

performance.

In both the flow cytometers, single cells are injected

through the flow cell and in front of a focus of a microscope

objective where each particle is hit by an excitation light

beam and the emitted signals are acquired. In the Bryte-HS

flow cytometer the incident light wavelength was selected in

the range 470–490 nm with a bandpass filter from a 150 W

Xe lamp. In the Apogee-A40 flow cytometer the incident light

wavelength is 488 nm produced by an Ar laser.

In both instruments, FALS signal (collected at small solid

angle, equal to 2� in the Bryte-HS), LALS (Large Angle Light

Scattering, collected at large solid angles) signal and 3 or 4

fluorescence signals were collected simultaneously for each

particle passing in front of the focus point. Green fluores-

cence emitted by SYBR-Green I was collected with a band-

pass filter with a bandwidth 515–565 nm in the Bryte-HS

flow cytometer and 515–545 nm in the Apogee-A40 flow

cytometer. During the flow cytometric analysis of bacteria

stained with SYBR-Green I, the threshold (for excluding

electronic noise) was set on the green fluorescence histo-

gram. Green fluorescence was used to gate FALS histogram,

in order to acquire only the FALS produced by nucleic acid-

stained bacteria and to eliminate scattering signals

originated by non-biotic particles or detritus present in

wastewater and activated sludge.

As expected, no difference was observed between the FALS

intensity of bacteria before and after the staining with SYBR-

Green I. This correspondence between the FALS intensity

without and with SYBR-Green I was verified with pure strains

in order to obtain a scattering signal well discriminated from

the instrument background.

FALS intensity of bacteria was recorded by adopting a linear

scale (that allows an easier conversion into biovolume with

respect to the logarithmic scale) mapped onto 256 channels.

A linear amplification factor from 10 to 40 was chosen depend-

ing on the size of the particles analysed.

For the analyses of silica beads the threshold was set only

on the FALS histogram, because beads are not fluorescent.

Every day, after daily calibration, during the analyses of differ-

ent beads or bacteria, no changes on photomultiplier values

were applied.

2.6. Modelling and theoretical aspects for describingFALS

The FALS intensity produced by bacterial cells was

explained on the basis of the Rayleigh–Gans approximation

of the Mie theory. The validity of using the Rayleigh–Gans

relationship for describing FALS of bacteria in wastewater

Fig. 1 – Optical configuration of the flow cytometer equipped with arc lamp and acquiring two scattering signals (at small

and large angles) and three fluorescences at different wavelength.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 7 5 7 – 3 7 6 63760

and activated sludge (but also in most environmental sam-

ples) is justified by two considerations minimizing the

phase shift:

(1) the size of bacteria generally ranges from 0.3 and 1 mm, not

too large with respect to the wavelength (l) of the excita-

tion light applied in the flow cytometer, of 488 nm in the

Apogee-A40 and in the range 470–490 nm in the Bryte-HS;

(2) bacteria are characterised by high water content, which

gives a refractive index similar to the surrounding me-

dium, which is usually water.

In fact in general the equation of Rayleigh–Gans can be

applied to particles with size between 0.3 and 3 mm and with

refractive index (indicated hereafter as n) similar to the one

of the medium (n0) surrounding the particle.

In flow cytometry the intensity of the scattered light

depends on the size (characterised by the radius, r, of a spher-

ical particle having an equivalent volume) and the orientation

of the particles (given by b angle) passing in front of the focus

point. The scattering light intensity (I ) at a certain angle of

observation (q) and distance (R) is (Koch et al., 1996):

I ¼ I0PðqÞ8p4r6n40

R2l4

"ðn=n0Þ2�1

ðn=n0Þ2þ2

#2

vy�1þ cos2q

�(2)

where P(q) for a rotation ellipsoid is:

PðqÞ ¼�3ðsin x� x cos xÞ

x3

�2

(3)

where the x term depends on the two axis (a,b) of the ellipsoid:

x ¼ 4pan0

lsin

�q

2

�"sin2

bþ�

ba

�2

cos2b

#12

(4)

The equations above allow the scattering intensity for any q

value to be calculated.

Summing up the Rayleigh–Gans equation, the FALS

intensity depends on:

- the bacterial properties, such as size or volume (related to r),

elongation (defined by the ratio b/a), orientation (b), and rel-

ative refractive index (n/n0);

- the optical characteristics of the flow cytometer, such as the

intensity of the incident light (I0), its wavelength (l) and the

angle and the distance of observation (q, R, respectively).

Furthermore, from Eq. (2), the scattering intensity is

proportional to the sixth power of r, which means it is propor-

tional to the square of the biovolume.

In order to investigate how the different parameters affect

FALS intensity, a sensitivity analysis of the model was carried

out.

The parameters of the flow cytometer (I0, l, R, q) are

characteristic of each instrument and do not change during

the analysis: therefore they can be assumed constant in the

model. The parameter l was fixed equal to 488 nm for simulat-

ing Apogee-A40, while l was assumed of 480 nm in the Bryte-

HS, as mean value of the range 470–490 nm (the sensitivity

analysis demonstrated that changes in FALS intensity due to

variations in l from 470 to 490 nm were less than 0.2% for par-

ticles as small as bacteria and so this influence is negligible).

The parameter q was assumed equal to 2�.

Among the parameters characteristic of bacterial cells (r, b/

a, n/n0 and b), the radius r (related to the volume) and the

relative refractive index, n/n0, gave the higher sensitivity as

discussed below, whereas the influence of elongation and

cell orientation is quite negligible.

Fig. 2 shows the relationship between the FALS intensity

predicted by the model (expressed in arbitrary units) and the

particle volume for varying values of the refractive index of

the particles and considering spherical particles. Values of

particle volume up to 0.6 mm3, similar to values of bacterial

Fig. 3 – Influence of particle elongation on the FALS

intensity predicted by the model, referring to two different

refractive indexes: (A) particles with n [ 1.38; (B) particles

with n [ 1.42.

Fig. 2 – Relationship between particle volume and FALS

intensity as predicted by the model for different values of

the refractive index of the particles.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 7 5 7 – 3 7 6 6 3761

cells, were considered. We selected refractive indexes, n,

typical of microbial cells, in the range 1.38–1.42. Water was

considered as surrounding medium (n0¼ 1.33).

By varying n a considerable difference in the FALS inten-

sity was observed. The higher the n value, the higher the

FALS intensity. Two particles having the same volume, but

different n (1.39 or 1.40), produce FALS intensities that differ

by 37%.

The influence of elongation of the particles (b/a ratio) on

FALS intensity predicted by the model is shown in Fig. 3.

Particles are considered as ellipsoids and elongation ranges

between 1 and 3 (1 in the case of spheres). The elongation

did not significantly affect the FALS intensity in the case of

particles made up of materials with n¼ 1.37–1.42, as indicated

in Fig. 3A and B where the curves coincide.

If FALS intensity would be simulated with a large angle of

observation (for example q higher than 20�), the influence of

elongation would be more evident (data not shown). In

conclusion, we can assume that elongation does not

significantly affect FALS intensity in the case of microbial

cells present in environmental samples (with n around

1.38–1.42).

On the basis of these results and considering particles in

the typical range size of bacteria, it can be concluded that

the FALS intensity is affected mainly by volume and refractive

index of bacterial cells and is approximately independent of

the shape of the cells.

3. Results and discussion

3.1. Calibration of FALS acquired by flow cytometry byusing silica beads as reference

The FALS intensity acquired by flow cytometry is expressed

in arbitrary units, usually applying a scale made up of 256 or

1024 channels. The purpose is to calculate the unknown ab-

solute volume of the bacterial cells starting from their FALS

intensity utilising reference beads having known size and

refractive index equal to bacteria. According to the model

explained above, the spherical shape of these reference

beads is adequate for representing also ellipsoidal bacterial

cells.

On the basis of these considerations, spherical beads made

of silica and not fluorescent were used as reference. Sizes sim-

ilar to bacteria were adopted, with beads volumes between

0.078 and 0.505 mm3. The refractive index of silica is equal to

1.42 (relative refractive index in water, n/n0, is equal to 1.065,

where n0 of water is equal to 1.333). This value is in the range

indicated in Section 1 for microorganisms in low-loaded

environments.

Since silica beads are not fluorescent, during the flow

cytometric analysis only the FALS signal was recorded and

represented in a histogram. The FALS signal produced by the

beads can easily be discriminated from electronic noise by

most flow cytometers on the market in the case of silica beads

Fig. 4 – Peak FALS channels acquired by flow cytometry as

a function of the volume of silica beads (square points) and

modelled FALS intensity predicted by the Rayleigh–Gans

theory: (A) data acquired by Bryte-HS; (B) data acquired by

Apogee-A40. Fig. 5 – Straight line for the conversion of FALS channels

(expressed in arbitrary scale) into particle volume: (A) data

acquired by Bryte-HS; (B) data acquired by Apogee-A40.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 7 5 7 – 3 7 6 63762

over 0.7 mm, whereas only some instruments are able to detect

the weak FALS signal produced by silica beads with smaller

diameters (lower than 0.5 mm). Using the Bryte-HS flow cytom-

eter it was quite difficult to distinguish from background

beads with diameter smaller than 0.7 mm on the basis of

FALS signal. The most recently developed Apogee-A40 flow

cytometer allowed to distinguish very clearly silica beads hav-

ing smaller diameters up to 0.4 mm. The low value of n of silica

contributes to cause the weak FALS signal of silica beads; on

contrary the most common polystyrene or latex beads, char-

acterised by higher n values, can be easily detectable also for

small diameters.

For each size of silica beads (six sizes of beads) the channel

corresponding to the peak of the FALS frequency distribution

(indicated hereafter as ‘‘peak FALS channel’’) was recorded

with both Bryte-HS and Apogee-A40 flow cytometers. The

data of peak FALS channels as a function of the correspondent

volume of beads are shown in Fig. 4 (square points) for both

the flow cytometers.

FALS intensity produced by spheres with n of 1.42 and

volume up to 0.6 mm3 was also calculated on the basis of the

Rayleigh–Gans model explained above (and indicated hereaf-

ter as ‘‘modelled FALS intensity’’) and is again shown in

Fig. 4. The modelled FALS intensity was used to evaluate the

congruence of the peak FALS channel obtained experimen-

tally with silica beads by flow cytometry. The good correspon-

dence between the two FALS intensities (Fig. 4) confirms the

effectiveness of the Rayleigh–Gans model in describing the

FALS signal acquired by both the flow cytometers. The perfor-

mance of the two flow cytometers – the Bryte-HS equipped

with arc lamp and Apogee-A40 equipped with laser – resulted

very similar. Therefore, the procedure used for the conversion

of FALS into biovolume is similar in both cases (see paragraph

‘‘Validation of the FALS conversion.’’). The relationship

between peak FALS channels and the second power of the

particle volume is well fitted with a straight line (R2> 0.98)

as indicated in Fig. 5.

This straight line can be used for the immediate evaluation

of biovolume of bacteria after measurement of FALS intensity

with flow cytometry. On the basis of this conversion, each

FALS channel is associated with a specific cellular volume

and the whole scale of FALS expressed in arbitrary units can

be converted in an absolute volume scale (expressing the

particle volume in mm3). From a practical point of view, the

Fig. 6 – Peak FALS channels of polystyrene beads as

a function of the volume (circle points) and modelled FALS

intensity predicted by the Rayleigh–Gans theory.

Fig. 7 – Cytometric analysis of bacteria in activated sludge

by using Bryte-HS: (A) frequency distribution of FALS

intensity; (B) frequency distribution of cellular biovolume

calculated from FALS intensity.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 7 5 7 – 3 7 6 6 3763

immediate conversion of FALS intensity into the correspond-

ing biovolume is based on the following equations:

Volume2 ¼ 0:0022 FALS� 0:0478

for the Bryte-HS flow cytometer (5)

Volume2 ¼ 0:0017 FALS� 0:0133

for the Apogee-A40 flow cytometer (6)

where FALS is expressed in channels, as acquired by the used

flow cytometer.

The intercept and the slope of the straight line (Eq. (5) or (6))

are coefficients depending on the specific characteristics of

the instrument used and therefore they could change for

different flow cytometers. Furthermore, the FALS intensity

produced by the same particles could change daily depending

on the setting of each instrument. As a consequence, Eq. (5) or

(6) may change and therefore the equation has to be verified

periodically in order to obtain accurate volume estimations.

The flow cytometric analysis of 4–6 different sizes of silica

beads and the calculation of Eq. (5) or (6) requires about 20–

30 min. Therefore, the calibration curve can be done easily

every day.

3.2. Inaccurate calibration of FALS by using polystyreneor latex beads as reference

As previously stated, inaccurate results would be expected, by

using beads made of fluorescent latex or polystyrene (widely

applied for daily calibration of cytometers) to simulate FALS

produced by bacterial cells. In fact, two particles having the

same shape and volume but significantly different refractive

indexes produce significantly different FALS intensities.

Polystyrene microspheres having diameter of 0.49, 0.87,

1.84, 2.27 mm and n of 1.585 were chosen. The FALS intensity

of beads was very strong, producing histograms well discrim-

inated from the electronic noise of the instrument. For each

size of beads, the channel corresponding to the peak of the

FALS frequency distribution was recorded (‘‘peak FALS chan-

nel’’) by using Bryte-HS flow cytometer and data are indicated

in Fig. 6 (circle points). Polystyrene beads suspended in water,

despite having size similar to bacteria, do not fall in the Ray-

leigh–Gans domain, as demonstrated by the large difference

from the modelled FALS intensity indicated in Fig. 6 as a con-

tinuous line.

3.3. Validation of the FALS conversion on bacteriapresent in wastewater and activated sludge

Bacterial cells present in wastewater and activated sludge

need to be stained with fluorescent probes, such as SYBR-

Green I, in order to be discriminated from other organic and

inorganic non-biotic particles, which can represent more

than 70% of the total solid concentration and could produce

the same FALS as bacteria. The scattering due to other non-

biotic particles was excluded from the analysis by gating

FALS on the bacteria population only. Before flow cytometric

Fig. 8 – Comparison between the bacterial biovolume

distribution measured under microscope and by Bryte-HS

flow cytometer (after conversion of FALS intensity) in the

case of (A) activated sludge and (B) wastewater.

Fig. 9 – Comparison between the bacterial biovolume

distribution measured under microscope and by Apogee-

A40 flow cytometer (after conversion of FALS intensity) in

the case of (A) activated sludge and (B) wastewater.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 7 5 7 – 3 7 6 63764

analysis, aggregates were dispersed by sonication in order to

obtain a suspension composed mainly of single cells.

The FALS distribution obtained for activated sludge

bacteria by using Bryte-HS flow cytometer is shown in

Fig. 7A. The typical arbitrary scale from 0 to 256 channels is

indicated on the horizontal axis. Only 200 channels were indi-

cated, because very few cells fell in the range 201–256 chan-

nels. To evaluate bacterial biovolume, FALS intensity was

converted on the basis of Eq. (5). The applicability of this con-

version is warranted by the following considerations:

- the elongation (b/a ratio) of bacteria in wastewater and

activated sludge was in the range between 1.5 and 2.5,

within the maximum frequency for ratios of 1.7. This evalu-

ation was done under epifluorescence microscopy after

fluorescent staining of cells with SYBR-Green I. For these

values of b/a ratio and bacterial biovolume lower than

1.0 mm3, FALS intensity is not significantly affected by elon-

gation, as demonstrated in Fig. 3 according to the model.

Therefore, the use of the calibration curve obtained with

spherical beads is correct;

- the influence of cell orientation (b) on FALS signal is

negligible, as evaluated by the model;

- the refractive index of silica beads should be similar to that

of bacteria.

As a consequence the FALS signal produced by bacteria is

related exclusively to their size or volume (related to r).

The frequency of biovolume distribution (expressed in

mm3), converted from the FALS intensity, is shown in Fig. 7B.

According to Eq. (5), channel 22 is the minimum FALS

channel that can be used, in order to avoid negative biovo-

lumes. Consequently, some (very small) particles emitting

FALS intensity lower than 22 are excluded from the conver-

sion. This minor fraction of small cells does not significantly

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 7 5 7 – 3 7 6 6 3765

affect the sizing of the entire bacterial population or the

subsequent calculation of cellular biomass.

The bacterial biovolume distribution calculated convert-

ing the FALS signal acquired by flow cytometry was com-

pared to the biovolume of the same bacteria directly

measured under epifluorescence microscopy using an auto-

mated image analysis (Fig. 8). The size classes used in flow

cytometry are smaller and consequently much more numer-

ous than in microscopy (about 180 against 20–30 classes).

The biovolume distribution obtained according to the two in-

dependent experimental approaches show a good similarity

both for bacteria in activated sludge (Fig. 8A) and wastewater

(Fig. 8B); these results are referred to the application of the

Bryte-HS flow cytometer.

Analogous procedure for the conversion of FALS intensity

was applied in the case of Apogee-A40 flow cytometer. The

obtained bacterial biovolume distributions are indicated in

Fig. 9A and B for bacteria in activated sludge and wastewater,

respectively. The samples shown in Fig. 9 are different from

those presented in Fig. 8 being characterised by bacteria

with smaller size. A good agreement was observed between

the biovolume measured under microscope and that calcu-

lated from FALS intensity acquired by the Apogee-A40

laser-based flow cytometer. Although the refractive index of

bacteria in wastewater and activated sludge is not known,

this good agreement between cytometric and microscopic

sizing confirms that the refractive index of bacteria in waste-

water and activated sludge is around 1.42. However, because

populations of cells are generally heterogeneous in regard to

their refractive indices, the value of 1.42 has to be assumed as

an average value.

The biovolume frequency distribution allows some

important parameters to be calculated: (1) total cellular bio-

volume present in a bacterial suspension (Vtot), by multiply-

ing each volume class by the number of cells falling in this

class; (2) mean cellular volume, by dividing Vtot by the total

number of cells; (3) peak volume, corresponding to the

most frequently occurring volume of bacterial cells; (4)

other frequency indicators such as median or percentiles.

Furthermore, the cellular biovolume can be converted into

bacterial biomass expressed as dry weight (for example

referring to VSS, Volatile Suspended Solids, or COD, Chem-

ical Oxygen Demand). VSS and COD are typical parameters

used in mass balance in WWTPs.

It can be observed from Fig. 8 that bacteria present in

wastewater are slightly greater (peak at 0.20 mm3) than the bio-

volume of cells in activated sludge (peak at 0.16 mm3), probably

due to the longer sludge age of bacteria in activated sludge

(about 12 days) with respect to raw wastewater coming from

the sewerage. The mean cellular biovolume (excluding

aggregates greater than 0.6 mm3) evaluated from the data

shown in Figs. 8 and 9 ranged from 0.14–0.32 mm3 for wastewa-

ter to 0.12–0.25 mm3 for activated sludge. Only particles with

biovolume smaller than 0.6 mm3 (corresponding to an equiva-

lent sphere diameter of 1.05 mm) could be considered as single

cells, while particles greater than 0.6 mm3 are attributed either

to large-size cells, such as eukaryotic organisms, or to

bacterial aggregates not dispersed during pretreatment by

sonication. In wastewater and activated sludge, the number

of particles having volume greater than 0.6 mm3 was less

than 12% with respect to the total number of cells counted

by flow cytometry.

Similar results were founded by other authors. Vollertsen

et al. (2001) referred mean and median values of 0.22 mm3 and

0.29 mm3, respectively, in raw wastewater, by sizing bacteria

under microscope. In activated sludge samples, Frølund

et al. (1996) measured a mean biovolume equal to

0.25� 0.24 mm3, while Jahn and Nielsen (1998) found cellular

biovolumes in the range of 0.13–0.37 mm3 for sewerage

biofilm.

4. Conclusions

The procedure for the rapid conversion of FALS acquired

with two models of flow cytometers (Bryte-HS and Apo-

gee-A40) into bacterial biovolume can be described as

follows:

(1) selecting 4–6 types of silica beads having known size and

relative refractive index similar to that of bacteria;

(2) building a calibration curve between the silica beads

volume and the correspondent FALS intensity (choosing

adequate PMT, gain and number of FALS channel);

(3) analysing the bacterial suspension with flow cytometry,

measuring the bacterial concentration and the FALS inten-

sity (applying the same PMT, gain and number of FALS

channels of point 2);

(4) converting the FALS intensity into bacterial biovolume

(each FALS channel can be immediately converted

into a correspondent volume) and finally plotting the

biovolume distribution for the bacteria present in the

sample.

The proposed procedure was validated in the specific case

of bacteria present in wastewater and activated sludge,

obtaining a good correspondence between the FALS conver-

sion and the sizing under microscope.

The main parameter affecting the FALS conversion is the

bacterial refractive index: for example by assuming a variation

of n from 1.40 to 1.42, an underestimation of 22% for biovo-

lume could occur.

Although the refractive index of bacteria in wastewater

and activated sludge is not known, the good agreement

between cytometric and microscopic size distributions con-

firms that the value of 1.42 may be a good estimation of the

refractive index of these bacteria.

Acknowledgements

The Authors are grateful to the reviewers for their precious

suggestions. The authors thank L. Bruni of the Biological and

Chemical Laboratories of the Autonomous Province of Trento

for her appreciated collaboration during the analytical stages.

S. Tamburini, M. Ferrai and F. Ierace of the Department of Civil

and Environmental Engineering of the University of Trento are

also gratefully acknowledged for their help in data analysis

and processing.

w a t e r r e s e a r c h 4 2 ( 2 0 0 8 ) 3 7 5 7 – 3 7 6 63766

r e f e r e n c e s

Bouvier, T., Troussellier, M., Anzil, A., Courties, C., Servais, P.,2001. Using light scatter signal to estimate bacterial biovolumeby flow cytometry. Cytometry 44, 188–194.

Davey, H.M., Davey, C.L., Kell, D.B., 1993. On the determination ofthe size of microbial cells using flow cytometry. In: Lloyd, D.(Ed.), Flow Cytometry in Microbiology. Springer-Verlag,London, pp. 49–65.

Davey, H.M., Kell, D.B., 1996. Flow cytometry and cell sortingof heterogeneous microbial populations: the importanceof single-cell analyses. Microbiological Reviews 60,641–696.

Foladori, P., Bruni, L., Andreottola, G., Ziglio, G., 2007. Effectsof sonication on bacteria viability in wastewater treatmentplants evaluated by flow cytometry – fecal indicators,wastewater and activated sludge. Water Research 41,235–243.

Frølund, B., Palmgren, R., Keiding, K., Nielsen, P.H., 1996.Extraction of extracellular polymers from activated sludgeusing a cation exchange resin. Water Research 30, 1749–1758.

Fry, J.C., 1990. Direct methods and biomass estimation. Methodsin Microbiology 22, 41–85.

Green, R.E., Sosik, H.M., Olson, R.J., DuRand, M.D., 2003. Flowcytometric determination of size and complex refractive indexfor marine particles: comparison with independent and bulkestimates. Applied Optics 42 (3), 526–541.

Jahn, A., Nielsen, P., 1998. Cell biomass and exopolymer compositionin sewer biofilms. Water Science and Technology 37, 17–24.

Jonasz, M., Fournier, G., Stramski, D., 1997. Photometricimmersion refractometry: a method for determining therefractive index of marine microbial particles from beamattenuation. Applied Optics 36 (18), 4214–4225.

Julia, O., Comas, J., Vives-Rego, J., 2000. Second-order functionsare the simplest correlations between flow cytometric lightscatter and bacterial diameter. Journal of MicrobiologicalMethods 40, 57–61.

Koch, A.L., Robertson, B.R., Button, D.K., 1996. Deduction of thecell volume and mass from forward scatter intensity ofbacteria analysed by flow cytometry. Journal ofMicrobiological Methods 27, 49–61.

Morel, A., Ahn, Y.H., 1990. Optical efficiency factors of free-livingmarine bacteria: influence of bacterioplankton upon the

optical properties and particulate organic carbon in oceanicwaters. Journal of Marine Research 48, 145–175.

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 and single cell sorting.Journal of Microbiological Methods 42, 97–114.

Porter, J., Deere, D., Hardman, M., Edwards, C., Pickup, R., 1997. Gowith the flow – use of flow cytometry in environmentalmicrobiology. FEMS Microbiology Ecology 24, 93–101.

Robertson, B.R., Button, D.K., Koch, A.L., 1998. Determination ofthe biomasses of small bacteria at low concentrations ina mixture of species with forward light scatter measurementsby flow cytometry. Applied and Environmental Microbiology64, 3900–3909.

Ross, K.F.A., 1957. The size of living bacteria. Quarterly Journal ofMicroscopical Science 98 (4), 435–454.

Salzman, G.C., Singham, S.B., Johnston, R.G., Bohren, C.F., 1990.Light scattering and cytometry. In: Melamed, M.R.,Mullaney, P.F., Mendelsohn, M.L. (Eds.), Flow Cytometry andSorting. John Wiley & Sons, Inc, New York, pp. 81–108.

Steen, H.B., 2000. Flow cytometry of bacteria: glimpses from thepast with a view to the future. Journal of MicrobiologicalMethods 42, 65–74.

Stramski, D., Bricaud, A., Morel, A., 2001. Modeling the inherentoptical properties of the ocean based on the detailedcomposition of the planktonic community. Applied Optics 40,2929–2945.

Twardowski, M.S., Boss, E., Macdonald, J.B., Pegau, W., Barnard, A.H., Zaneveld, J.R.V., 2001. A model for estimating bulkrefractive index from the optical backscattering ratio and theimplications for understanding particle composition in case Iand case II waters. Journal Of Geophysical Research 106 (C7),14129–14142.

Valkenburg, J.A.C., Woldringh, C.L., 1984. Phase separationbetween nucleoid and cytoplasm in Escherichia coli as definedby immersive refractometry. Journal of Bacteriology 160,1151–1157.

Vollertsen, J., Jahn, A., Nielsen, J.L., Hvitved-Jacobsen, T.,Nielsen, P.H., 2001. Comparison of methods for determinationof microbial biomass in wastewater. Water Research 35,1649–1658.

Ziglio, G., Andreottola, G., Barbesti, S., Boschetti, G., Bruni, L.,Foladori, P., Villa, R., 2002. Assessment of activated sludgeviability with flow cytometry. Water Research 36, 460–468.