Recent Sensing Technologies For Pathogen Detection in Milk A Review

Biosensors and Bioelectronics 60 (2014) 821
Contents lists available at ScienceDirect
Biosensors and Bioelectronics

journal homepage: www.elsevier.com/locate/bios
Recent sensing technologies for pathogen detection in milk: A review

Alessia Mortari n, Leandro Lorenzelli
BioMEMS, Center for Materials & Microsystems, Fondazione Bruno Kessler, Via Sommarive 18, 38123 Trento, Italy
art ic l e i nf o
a b s t r a c t
Article history:
Received 25 November 2013
Received in revised form
7 March 2014
Accepted 26 March 2014
Available online 12 April 2014
Quality control utilising Hazard Analysis and Critical Control Points in the dairy industry generates a
large volume of samples. The associated costs are signicant. The development and application of fast,
sensitive and cost effective analytical systems for pathogen detection in milk could aid the industry in
the reduction of overheads, nd new uses in dairy farming and production precision management and
unlock new markets. Recent progress in pathogen sensing technologies for milk analysis, in particular
nucleic acid amplication and biosensors, is reviewed here. The importance of representative samples,
detection probability and Practical Detection Limit is claried. Methods for sample pretreatment are
discussed in association with the most applicable detection methods. The major ndings are summarised and future perspectives are drawn to inspire new ideas in the scientic community.
& 2014 Elsevier B.V. All rights reserved.
Keywords:
Pathogen detection
Milk
Detection probability
Lab on chip
Contents
1.
2.
3.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Milk sampling and sample preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Methods for foodborne pathogen detection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.1.
Nucleic acid amplication based technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1.1.
Sample preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1.2.
DNA extraction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1.3.
PCR variants and amplicons analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1.4.
Loop mediated isothermal amplication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.1.5.
Nucleic acid sequence based amplication. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2.
Biosensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2.1.
Electrochemical biosensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.2.2.
Optical based biosensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.2.3.
Mass sensitive biosensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.3.
Other sensing technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.3.1.
Flow cytometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.3.2.
Impedance microbiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.3.3.
Spectrouorometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.3.4.
Multisensory systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
4. Lab on chip . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5. Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
6. Future perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Acknowledgement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1. Introduction
n
Corresponding author. Tel.: 39 046 131 4132; fax: 39 046 130 2040.
E-mail addresses: mortari@fbk.eu, mor_ale@hotmail.com (A. Mortari).
http://dx.doi.org/10.1016/j.bios.2014.03.063
0956-5663/& 2014 Elsevier B.V. All rights reserved.
Foodborne pathogen testing has been regulated with the introduction of Hazard Analysis and Critical Control Points (HACCPs)
A. Mortari, L. Lorenzelli / Biosensors and Bioelectronics 60 (2014) 821
100
Number of publications
(EC Health & Consumer Protection General Directorate, 2005; FDA,

1997). Based on risk analysis along the production and distribution
chain, HACCP promotes the identication of critical points for
sampling and testing. Its aim is to ensure food safety. As a parallel
result, the testing volume has increased and shifted toward pathogen
detection on-site at production plants (Alocilja and Radke, 2003).
The dairy sector has been agged with the highest testing rate per
processing plant, but despite quality control, refrigeration, pasteurisation and ultra-high temperature (UHT) treatments, outbreaks of
foodborne illnesses due to the consumption of hygenised but
contaminated dairy products have been reported (Ackers et al.,
2000; De Buyser et al., 2001; Morgan et al., 2009; Schmid et al.,
2009; Upton and Coia, 1994).
The dairy industry is in need for a fast, sensitive and cost
effective technology for the detection of foodborne pathogens that
could successfully address the tasks appointed by legislator bodies.
From the FoodMicroSystems (EU FP7/2007-2013 project, grant
agreement no. 287634) and its analysis on the dairy industry
(FoodMicroSystems, 2013) pathogen detection has emerged as the
highest technological priority for the dairy industry. The industry
demand for new technologies is mainly motivated by nancial
considerations in relation to dairy farming, quality control and
production precision management, such as efcient use of resources
and waste reduction, safety and better use of raw milk that could
facilitate commercialisation of new and diversied products. Consumer perspective is mainly driven by cost and safety but also by
quality and health. Regarding these factors, the application of
innovative and advanced analytical systems for pathogen detection
in raw milk is a good example where these technologies could
unlock new markets and products. Raw milk and raw milk products
are experiencing an increasing worldwide market demand (Food
Standards Australia New Zealand, 2009; www.realmilk.com; www.
milkmaps.com). Consumers' demand for raw milk access is based
on scientic evidence that raw milk has superior nutritional
properties (Quigley et al., 2013; Debarry et al., 2007; Ege et al.,
2012). In conjunction, farmers are driven by the will to deliver new
high quality products and the interest to grow sustainable small and
medium enterprises (EC Agriculture and Rural Development, 2005;
Farm-to-Consumer Legal Defence Fund, 2013). However, raw milk is
a potential vehicle for outbreaks of foodborne illnesses (LeJeune and
Rajala-Schultz, 2009) and the medical community warns against
the risks posed by unpasteurised raw milk. Local legislations have
been developed to address the consumer demand whilst providing
standards for the production of raw milk (EC, 2004) or alternatively,
to ban the trade of raw milk for human consumption (Farm-toConsumer Legal Defence Fund, 2013; Food Standards Australia
New Zealand, 2012). The reasoning behind the legislation is based
on the fact that raw milk cannot be certied safe within its life
cycle: raw milk is a highly perishable food whilst pathogen analyses
are time consuming. As consequence, the microbial analysis of raw
milk is merely retrospective, has statistical value but cannot prevent
a disease outbreak. In addition, the available methods are excessively expensive to apply to the certication of every raw milk
consignment.
A literature analysis carried out with ISI Web of Knowledge
using the keywords pathogen detection milk has underlined an
increased scientic interest in the topic (Fig. 1). The interest may
well be in response to the enforcement of HACCP regulation.
When the analyses is rened with the keywords integrated,
automated, sensor and system the trend growth is conrmed
and the commitment of the scientic community to the development of analytical systems for accurate quality control and
effective detection of pathogens in milk is underlined.
This review aims to cover the topic of sensing technology for
pathogen detection in milk and the requirements for their application at an industrial level. It provides direction on research and
10
1
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Year
Key words: "pathogen detection in milk"
Refined by: "integrated", "automated", "sensor" and "system"
Fig. 1. Number of publications in logarithmic scale since 2000 till 2012. The
publications were collected and analysed using ISI Web of Knowledge.
development for advanced and integrated analytical systems that

could full the needs of the dairy industry. After an initial discussion of milk sampling and sample preparation, a critical review of
the most applicable technologies developed since 2000 is carried
out. The main ndings are summarised and future perspectives are
drawn to inspire new ideas in the scientic community.
2. Milk sampling and sample preparation

Milk is a complex matrix (Fox, 2011). It is a colloid of butterfat
globules and water with dissolved carbohydrates and protein
complexes. Bacteria are distributed throughout the emulsion:
suspended in solution as well as entrapped and absorbed on
proteins micelles and fat globules. Milk is highly nutritious, with
neutral pH and high water activity, thus it can support the growth
of a rich and complex microbial ora. Produced from a healthy
udder, milk is sterile. During the descent and excretion through
the teat it is colonised by health promoting and technologically
important bacteria, such as Lactobacillus spp. Other sources of
potential contamination are personnel and milking, transport,
storage and processing equipment. Zoonotic pathogens can be
introduced by mastitic and unhealthy animals, bedding material,
faeces and soil. The most common pathogenic bacteria associated
with milk are Listeria spp., Salmonella spp., Escherichia coli,
Campylobacter spp., Shigella spp. and Brucella spp. (Quigley et al.,
2013). Table 1 reports the acceptable pathogen concentration for
microbial analysis of milk as specied in the Commission Regulation (EC) No. 2073/2005 (EC, 2005) along with the respective
reference methods, sample volumes and analysis time.
As specied in the guidelines ISO 707:2008 Milk and milk
products: guidance on sampling (ISO, 2008), a sample of milk
has to be representative, thus be equivalent to the whole milk
consignment composition. Its weight is of 25 g, corresponding
approximately to 25 mL of milk. Sample volume is of particular
importance in relation to the probability of sampling pathogenic
bacteria, normally dened as Detection Probability and expressed
in % value. Pathogens are normally subdominant species and their
infection dose is minimal, in the order of 10 and 100 CFU/mL
(Dineen et al., 1998). Zero tolerance is applied to particularly
virulent microorganisms, such as Listeria monocytogenes, Salmonella, E. coli O157:H7 and Enterobacteriaceae (EC, 2005; FDA, 2012).
In such cases, pathogen absence has to be certied through
microbiological analysis and the analysis sensitivity must be of
at least 1 CFU/10 g.
10
Table 1
Pathogens acceptable concentration as specied in the Commission Regulation (EC) No. 2073/2005 (EC, 2005).
Micro-organism
Food category
Acceptable concentration
Analytical reference method
Analysis time, days
Enterobacteriaceae
Pasteurised milk
Milk powder
Raw milk
Ready-to-eat food
Raw milk
Milk powder
Milk powder
o 1 CFU/mL
o 10 CFU/g
o 10 CFU/g
Absence in 25 g
Absence in 25 g
ISO 21528-1
6.57.5
ISO 16649-1 or 2
EN/ISO 11290
EN/ISO 6579
1
35
4.5
o 10 CFU/g
EN/ISO 6888-1 or 2
2.5
Escherichia coli
Listeria monocytogenes
Salmonella
Coagulase-positive Staphylococci
Table 2
Commercially available pathogen sensing systems.
Company
Product name
Technology
Application
Website
Life Technologies
Veredus Laboratory Pte. Ltd.
nanoRETE Inc.
BioDetection Instruments
7500 PCR System

VereFoodborne
X-MARK
Aegis 1000
Laboratory
Laboratory
Field use
Laboratory
www.lifetechnologies.com
www.vereduslabs.com
www.nanorete.com
www.biodetection-instruments.com
BIACORE
Sensata Technologies Inc.
Innovative Biosensor Environment
Group Inc.
Agilent Technologies
BIACORE Q
Spreeta
BioFlashs
qPCR
Lab on chip and mPCR
Antibody functionalised nanoparticle biosensor
Bio separator/antibody based optical/
electrochemical biosensor
SPR biosensor
SPR biosensor
Antibody based chemiluminescent biosensor
Laboratory
Laboratory
Laboratory
www.biacore.com
www.sensata.com
www.innovativebiosensors.com
On chip ow cytometry
Laboratory
www.genomics.agilent.com
Agilent 2100
Bioanalyzer
Sample preparation of standard methods includes a preliminary step of sample enrichment in broth culture, which normally
takes place overnight. The aim is to increase the pathogen concentration to a detectable level but the enrichment step is time
consuming. Furthermore, stressed cells may not multiply, hence
it may not be possible to establish a quantitative relation between
the initial sample and the analysed sample that is produced after
enrichment. The detection of viable but not culturable cells has
become a difcult task of interest to the scientic community
(Oliver, 2005). Advanced sensing technologies are required to
analyse representative samples in short time frames and produce
sensitive results. However, they have been mostly developed to
process small sample volumes, often in the order of hundreds mL.
Whilst detection technology is moving fast toward the reduction
of analysis time, the biggest challenge still remains the time and
labour effort required for sample preparation (FoodMicroSystems,
2013). Bacteria separation from the matrix and its concentration
have been recently applied to the reduction of sample volume to
an adequate size (Stevens and Jaykus, 2004) and to increase the
method sensitivity. In order to release adsorbed bacteria from the
milk matrix, methods including biochemical techniques based on
enzymatic reactions, surfactants and detergents have been developed. Matrix separation and pathogen concentration have been
accomplished with centrifugation, ltration and immunomagnetic
separation. Bacteria concentration and matrix removal on 3D
structures, such as pillars or herringbone, via bio-selective molecules, are also possible. Such methods have the potential to be
utilised in compact, automated and integrated systems. Still, milk
is a complex and challenging matrix that is well known for protein
fouling (Chemburu et al., 2005; Laczka et al., 2011), inhibition of
biochemical reactions (Chen et al., 2008; Rossen et al., 1992) and
membrane blockage (Chemburu et al., 2005).
Commercial solutions are available. Innosieve Diagnostic BV
manufactures a ltering system for laboratory applications that
signicantly decreases time to result and BioDetection Instruments produces a bioreactor for sample preparation integrated
with a biosensor for detection (Table 2).
An in-depth discussion on the topic of matrix separation and
bacteria concentration from milk is articulated along with and in
Impedance
microbiology
3%
e-sensors
5%
MS
5%
Flow
cytometry
11%
Biosensors
28%
Raman
spectroscopy
3%
Nucleic acid
amplification based
technologies
45%
Fig. 2. Publication percentiles on sensing methods for the detection of foodborne

pathogens in milk since 2000 (MS: mass spectrometry).
relation to each technical section of this review. Yet, none of the

developed methods have proven ideal and further research is
required to solve the issue of milk matrix inhibitory affects and
sample concentration.
3. Methods for foodborne pathogen detection

Conventional methods for qualitative and quantitative detection of pathogens rely on media growth. Pre-enrichment in nonselective culture broth, selective enrichment and plating, biochemical screening and serological conrmation are the basic steps
involved in the classical methods. Those are time consuming
procedures that require days for conrmation, are labour intensive
and rely on interpretation by skilled personnel interpreter
(Mandal et al., 2011).
Recently developed methods are based on nucleic acid amplication, biosensors, ow cytometry, spectrometry techniques and
multi-sensors systems. Fig. 2 represents the variety of technologies
11
Table 3
Selected research articles on the topic of nucleic acid amplication technologies in milk microbiology. Abbreviations: D.L.: detection limit; L.R.: linear range; A.T.: analysis
time; Sp.: specicity, compared to conventional culture methods; D.P.: detection probability.
Microorganism
Target gene
Sample type and

volume
End point PCR

L. monocytogenes
hlyA
Pasteurised whole
milk
10 mL
M. avium
paratuberculosis
IS900
Real Time Quantitative PCR

C. jejuni
VS1
Sample preparation and

DNA isolation
Performance evaluation
Reference
(I) Matrix separation by centrifugation

(II) In house cell lysis method
(III) Magnetic beads based DNA isolation
D.L.: 10 CFU/mL
Amagliani et al. (2006,

2004)
Raw milk
50 mL

(II) Bacteria immunomagnetic separation
(III) In-house DNA extraction method
D.L.: 0.1 CFU/mL

A.T.: 2 days
Djnne et al. (2003)
Raw milk
50 g

(II) DNA extraction with kit (Promega)
D.L.: 6-15 CFU/PCR tube

A.T.: 1 h
Yang et al. (2003)
D.L.: 1 CFU/mL
A.T.: 3 h
Paul et al. (2013)
D.L.: 10 CFU/mL
A.T.: 24 h
Khare et al. (2004)
D.L.: 5 CFU/mL
Sp.: 99.6% inclusivity
100% Exclusivity
Hein et al. (2006)
A.T.: 7 h
E. coli O157
rfb
Raw milk
10 mL
(I) Sample preparation using EDTA and

matrix
separation by centrifugation
(II) In house DNA extraction method
M. avium
paratuberculosis
IS900
Raw milk
2 mL
Salmonella spp.
invA
Raw milk
25 mL
Matrix separation by centrifugation

Bacteria immunomagnetic separation
Beads beating
In-house DNA extraction method
(I) Enrichment culture
(II) DNA extraction with kit (NucleoSpin
Tissue)
Raw milk
50 mL

(II) Matrix separation by centrifugation
(III) In-house DNA extraction method
D.L.: 100 CFU/mL
Giannino et al. (2009)
S. aureus
Y. enterocolitica
14 Lactic acid
bacteria
E. coli O157
L. monocytogenis
Salmonella spp.
hipO and mapA

stx1 and stx2
plcA, plcB and
inlB
rfbJ, iC, iB and
invA
nuc and lukE
ail and yst
RfbE
hlyA
ttr
Raw milk
125 mL

(II) DNA capture with magnetic
nanoparticles or
DNA extraction with column based kit
(Genomic DNA From Bacteria in Milk)
D.L.: 1 CFU/125 mL
A.T.: 2 days
Sp.: 100% inclusive 100%
exclusive
Omiccioli et al. (2009)
LAMP
L. monocytogenis
iap
Raw milk
1 mL
D.L.: 186 CFU/mL
Wang et al. (2010)
phoP
Raw milk
25 mL
Enrichment culture
Sample preparation with EDTA, surfactant
and dilution
In-house DNA extraction method
(II) Matrix separation by centrifugation
(III) Cell lysate by boiling
(IV) Debris separation by centrifugation
D.P.: 100%
D.L.: 35 CFU/250 mL
A.T.: 24 h
Li et al. (2009)
tmRNA/ssrA
Raw milk
Sample preparation using EDTA, detergent

and latex beads
Matrix separation by centrifugation
Kit based RNA extraction (RiboPure
Yeast Kit)
D.L.: 10 CFU/mL
O'Grady et al. (2009a)
Multiplex PCR
C. jejuni
E. coli O157:H7
L. monocytogenes
S. typhimurium
Salmonella
NASBA
S. aureus
1 mL
developed in this century for the detection of pathogens in milk.

Nucleic acid amplication based technologies have engaged
almost half of the research effort while biosensors represent the
second largest big sector of development. Nucleic acid amplication technologies and biosensors hold most of the expectation due
to their high sensitivity and selectivity.
A.T.: 34 h
The authors should point out that in many of the reviewed

articles a discrepancy exists between the method and its reported
detection limit unit. The most common unit is CFU/mL which
implies the quantication of Colony Forming Units, thus the
detection of viable cells only. However, CFU/mL was used to
quantify detection limit of total analysis that includes viable,
12
viable but not culturable and dead cells (e.g. Djnne et al., 2003;
Kim et al., 2010). If the method does not include differentiation
between viable and dead cells, the correct unit should be cells/mL.
In this review paper, the original unit used in the cited research
articles for detection limit characterisation has been reported. The
topic of differential detection between viable and dead cells is of
high relevance to the development of technologies for foodborne
pathogen detection and has been discussed in this review in
relation to each method reviewed.
3.1. Nucleic acid amplication based technologies
Various molecular methods based on enzymatic amplication
of specic nucleic acid segments have been applied to the detection of pathogens in milk. The technologies discussed hereby
are Polymerase Chain Reaction (PCR), Loop Mediated Isothermal
Amplication (LAMP) and Nucleic Acid Sequence Based Amplication (NASBA). Nucleic acid amplication based technologies
include four main steps: (i) sample preparation, (ii) DNA extraction, (iii) nucleic acid amplication and (iv) amplicons analysis.
3.1.1. Sample preparation

Sample preparation presents important challenges. In particular, sample volume is relevant in relation to its representability
and the nal amplication reaction volume, as well as the method
sensitivity itself (see Paragraph 2). As described in the guidelines
ISO 707:2008 Milk and milk products: guidance on sampling
(ISO, 2008), representative sample must be of 25 g. Amplication
reaction tubes have a total volume of 200 mL, which allows for a
maximum sample volume of about 100 mL. It is clear that the rst
challenge is to concentrate the sample. The concentrated sample
must be of high purity, thus free from inhibitors and with a
minimum necessary template concentration to produce detectable
amplicons.
Representative samples are often inoculated in broth culture
for overnight sample enrichment (Perelle et al., 2004; Wang et al.,
2010). The overnight enrichment allows achievement of high
sensitivity although viable but not culturable cells are of impediment to the correlation of the initial sample concentration and the
nal amplicons enumeration (Vanegas et al., 2009).
Centrifugation is another common approach to sample preconcentration. Centrifugation, thus bacteria pelleting, is also used
to separate matrix, such as broth culture or milk, from microorganisms (Djnne et al., 2003; Paul et al., 2013). Milk matrix
separation is necessary in nucleic acid amplication techniques
since milk is a complex matrix, well known for its inhibitory
inuence on DNA extraction and PCR (Ongol et al., 2009; Rossen
et al., 1992).
Immunomagnetic separation may improve bacteria isolation
from the sample matrix (Djnne et al., 2003; Hsih and Tsen, 2001).
Djnne et al. succeeded in achieving the lowest detection limit
reported in PCR literature by applying immunomagnetic separation and dot blot amplicons detection to the analyses of Mycobacterium avium subsp. paratuberculosis. The detection limit was of
0.1 CFU/mL from a 50 mL sample of raw milk.
3.1.2. DNA extraction

The DNA extraction must be of high efciency in order to
achieve the highest sensitivity of amplicon detection. Furthermore, an insufcient recovery of DNA template may be the origin
of false negative results. Traditionally, DNA extraction is obtained
with the use of toxic organic solvents (Giannino et al., 2009; Khare
et al., 2004). DNA extraction from milk samples with thermal
shock (Li et al., 2009; Rossmanith et al., 2006) and enzymatic
methods (Amagliani et al., 2006, 2004; Djnne et al., 2003) is also

possible. Currently, methods based on commercial kits are preferred (Hein et al., 2006; Yang et al., 2003). These methods are
time consuming and generally difcult to automate. However,
progress toward robotic DNA extraction has been made. Ehrs et al.
utilised a robotic system for the automated DNA extraction with
three commercially available methods based on magnetic beads
and silica gel ltration that was successfully applied to milk
samples (Ehrs et al., 2011). Life Technologies Pathatrixs and
Prep SEQs are two automated systems for sample preparation.
DNA purication utilising magnetic bead technology was proposed as an alternative to address the above listed limitations
(Amagliani et al., 2006, 2004; Omiccioli et al., 2009). The utilisation of complementary oligonucleotide probes linked to magnetic
nanoparticles simplies the DNA extraction process and could
replace the need for sample enrichment. Magnetic bead technology for DNA purication was suggested as being suitable for
automation, thus minimising labour and analysis time. Magnetic
beads produce reproducible yield results and are applicable to
large scale analysis.
The topic of total extraction was addressed by Riyaz-Ul-Hassan
et al. (2013) who developed a DNA extraction method aiming
to remove the fat component of milk. Directly from full fat milk,
Salmonella cells were rst precipitated with ammonium sulphate,
washed and the DNA extracted. The authors' ndings proved that
the preliminary step with ammonium sulphate produced a higher
DNA concentration yield than the untreated samples. The extraction method took 5 h.
An in-house method for DNA extraction was developed by
Cremonesi et al. (2006). The method was based on chemical lyses
and enzymatic inactivation combined with silica particles functionalised with nucleic acid binding properties. It was applied to the
DNA extraction of Gram positive mastitis pathogens directly from
milk. The authors claimed that the procedure could be automated
with the implementation of a liquid handling system and that high
throughput sample processing could be achieved, providing reductions in extraction time and cost.
Strotman et al. (2012) applied immiscible ltration assisted by
surface tension (IFAST) to bacterial DNA extraction and purication
from milk samples. IFAST technology exploits the unique properties of microuidics where surface tension dominates over gravity
and generates virtual separation walls between immiscible phases.
The two phases work as a lter to separate pure DNA extract from
the matrix. IFAST was suggested as an alternative to wash and
centrifugation steps typical of DNA solid phase extraction methods. It was applied to the DNA extraction of Clostridium botulinum
from milk and other food matrices. The method was compared to a
commercial kit and proved equivalent, but the sample volume was
limited to 8.5 mL.
3.1.3. PCR variants and amplicons analysis

Starting from the pioneering work of Mullis (Mullis et al., 1987),
PCR has experienced a continuous growth in research elds
and has become a valued technology in biomolecular laboratories
for various applications, including the microbiological analysis of
milk. ISO standards are available as guidelines for the qualitative
detection of foodborne pathogens (ISO, 2005a, 2005b, 2006a,
2006b, 2011). Commercial PCR systems are available by Life
Technologies and Veredus Laboratory Pte Ltd. The later product
is VereFoodborne, a lab on chip, which includes multiplex PCR
and microarray hybridisation. Life Technologies products include
automated robotic sample preparation, assay reagents and instrumentation. A selection of the reviewed research papers on the
topic of PCR microbiology in milk analysis is summarised in
Table 3.
PCR is highly selective and is able to detect subdominant

species. For instance, Nam et al. (2005) reported 100% inclusivity
and 100% exclusivity for the detection of Salmonella spp. in raw
milk samples. The target segment was of 119 base pairs and was
obtained from the invA gene. 124 Salmonella spp. and 116 nonSalmonella spp. were tested for the evaluation.
PCR variants and amplicon analysis methods have been developed. The classic end point PCR (epPCR) and its developments
such as quantitative or real time PCR (qPCR), multiplex PCR
(mPCR), reverse transcriptase PCR (rtPCR), viability PCR (vPCR)
and their application in the detection of pathogens in milk are
discussed here.
epPCR amplicon products are usually conrmed on gel electrophoresis with the use of DNA intercalating agents, such as ethidium
bromide. Djnne et al. (2003) applied epPCR for the DNA amplication
of M. avium subsp. Paratuberculosis in 50 mL raw milk samples. The
amplicons were analysed in gel electrophoresis yielding a detection
limit of 1 CFU/mL and with dot blot technique achieving the detection
limit of 0.1 CFU/mL. epPCR was used in combination with DNA
magnetic capture of L. monocytogenes in pasteurised whole milk
(Amagliani et al., 2006, 2004). It was reported that the passivation of
the magnetic particles with blocking agents was fundamental in
eliminating PCR inhibitory properties. An alternative multiplex
mesouidic system based on uorescent beads for the detection
of eight foodborne pathogens in milk powder was published by
Jin et al. (2009). Amino-modied oligonucleotide probes were immobilised on the beads and placed in the channels. A mixture of different
pathogen PCR products was injected through the channels for a
30 min hybridisation reaction. The chip was then scanned for spot
analysis and quantication. The detection limits were in the order of
103 CFU/mL. Despite the high detection limit, the system was fast and
able to analyse multiple samples simultaneously. epPCR was also
associated with amplicon detection in sandwich hybridisation assays
(Daly et al., 2002; Perelle et al., 2004) with good results. Daly et al.
analysed E. coli in 100 mL of pasteurised milk and obtained a detection
limit of 5 CFU/mL. The analyses took over 1 h to complete. The
authors argued that the proposed PCR ELISA method was fast,
straightforward, inexpensive and, in comparison to gel electrophoresis, more prone to automation and multiple sample analysis.
Furthermore PCR ELISA excludes the subjectivity of band identication and the use of carcinogenic DNA binding agents.
qPCR is an automated method where uorescent reporter
detection enables amplicon quantication in real time. The risk
of cross contamination and labour effort are minimised. qPCR has
found wide applications in pathogen analysis in milk (Paul et al.,
2013; Yang et al., 2003). For instance, Paul et al. applied qPCR to
the detection of E. coli O157 to 10 mL of raw milk and a 1 CFU/mL
detection limit was obtained within 3 h of total assay time.
mPCR has also been applied to the multiple and simultaneous
detection of foodborne pathogens in milk (Hsih and Tsen, 2001;
Omiccioli et al., 2009). Omiccioli et al. applied mPCR in the analysis
of raw milk for the simultaneous detection of E. coli O157,
L. monocytogenis and Salmonella spp. After overnight sample enrichment the authors achieved a detection limit of 1 CFU/125 mL,
100% inclusivity and 100% exclusivity. Lee et al. (2008) published
on a biochip for mPCR amplicons detection of mastitis causing
pathogens in UHT milk. The biochip was based on DNA hybridisation and a colorimetric reaction where probes for seven
pathogens were spotted on specic positions of the microarray
polymer substrate. The recorded detection limits were in the order
of 103 and 105 CFU/mL but the articially contaminated milk
samples volume was of 140 mL, which may have been partly the
origin of the poor sensitivity. The DNA hybridisation reaction took
140 min, the colorimetric reaction 60 min and the whole procedure, from DNA extraction to amplicon quantication, required a
total time of 6 h.
13
rtPCR represents a step forward in the detection of viable cells.

rtPCR targets RNA transcripts, therefore can provide information
on microbial activity. RNA extraction is more difcult than DNA
extraction, especially from complex and fatty matrices (GarcaNogales et al., 2010; Ulve et al., 2008). RNA is subject to fast
degradation and must be analysed quickly. Furthermore, RNA copy
number is variable and depends upon various factors that may
inuence cell growth phase and activity. Therefore, rtPCR is not
suggested for quantitative analysis. Only two publications were
found on the application of rtPCR to microbial detection in milk
(Choi and Lee, 2004; Vaitilingom et al., 1998).
vPCR has also been developed to discriminate between viable
and dead cells. Prior to the DNA extraction, free DNA that had been
released from dead cells is blocked with an impermeable nucleic
acid binding dye, which will prevent the amplication of such DNA
during the PCR. No publications on vPCR in milk pathogen detection
were found, but some studies were carried out utilising fermented
milk and yogurt (Falentin et al., 2010; Sheu et al., 2010, 2009).
For further reading on the subject of PCR in food and dairy
microbiology, we recommend the reviews written by Boyer and
Combrisson (2013) and Postollec et al. (2011).
3.1.4. Loop mediated isothermal amplication
A direct competitor to PCR is LAMP. LAMP is a technique
recently developed by Notomi et al. (Notomi et al., 2000; Notomi
and Hase, 2002). Based on DNA amplication under isothermal
conditions, LAMP is highly specic as six different regions of a
gene are targeted by independent sequences for a selective
identication. Later, four different primers are used in combination to amplify the target gene segment, thus offering higher
sensitivity and shorter analysis time.
LAMP has been applied to the detection of various foodborne
pathogens in milk (Liu et al., 2007; Duarte et al., 2013). A selection
of the reviewed research papers in the topic of LAMP microbiology
in milk analysis is summarised in Table 3.
Wang et al. (2010) published a comparison study between
LAMP and PCR for the analysis of L. monocytogenis in raw milk.
The detection limit for LAMP was 186 CFU/mL whereas for PCR it
was three orders of magnitude greater. The LAMP amplication
time was 40 min. Li et al. (2009, 2010) applied LAMP to the
detection of Salmonella and Yersinia enterocolitica from 25 mL of
raw milk and 100 g of powdered milk respectively. The reported
detection limits were of 35 CFU/250 mL for Salmonella and
2.2 CFU/100 g for Y. enterocolitica. The authors also calculated a
detection probability of 100% for both methods when the DNA was
extracted from 101 CFU.
3.1.5. Nucleic acid sequence based amplication
NASBA is based on the enzymatic activity of reverse transcriptase that amplies RNA templates into complementary DNA.
Developed by Compton (1991), NASBA operates under isothermal
conditions and with real time rapid detection of amplicons. As per
rtPCR, NASBA presents the advantage of viable cells detection.
Furthermore, NASBA operates in isothermal condition and, as per
LAMP, holds the promise of simple integration into a compact
system.
A publication was found for pathogen analyses in milk, where
O'Grady et al. (2009) applied NASBA to the amplication of tmRNA
for the detection of Staphylococcus aureus in 1 mL raw milk. The
method proved highly sensitive, with a detection limit of 1 CFU/mL
without the need for pre-enrichment.
3.2. Biosensors
Biosensors are characterised by intrinsic selectivity that allows
direct analysis of complex samples with minimal or no sample
14
purication. The most common transducers in milk analysis are

electrochemical, followed by optical and mass sensitive. Biosensor
transducers hold the promise to be relatively cheap and portable.
The signal is generated in real time, making biosensors fast analytical
devices. Thanks to their exibility in design and conguration,
biosensor technology has embraced the advancements offered by
micro- and nanotechnology, which have been applied to analyte
capture, signal amplication and sensor fabrication (Prez-Lpez and
Merkoi, 2011). The potential of biosensors has raised high expectation but their use is still mostly limited to research laboratories with
only a few biosensors present in the market. Current commercial
biosensors for foodborne pathogen detection are X-MARK by
nanoRETE Inc. and Aegis 1000 by BioDetection Instruments.
A selection of publications is summarised in Table 4. The
reviews written by Velusamy et al. (2010) and Lazcka et al.
(2007) are suggested for further readings on the wider subject of
biosensors for foodborne pathogens.
3.2.1. Electrochemical biosensors

Electrochemical biosensors are common in the literature. Electrochemical transducers are inexpensive and are not affected by sample
turbidity, such as that presented by milk. Furthermore, the resulting
electrical signal is directly converted in the electronic domain,
allowing the development of simple instrumentation and compact
system design.
Amperometric biosensors are often associated to an indirect
pathogen detection via an enzyme-linked immunoassay sandwich.
In a classic design, primary antibodies are immobilised on the
working electro-surface with a mediator for efcient electron transfer (Escamilla-Gmez et al., 2008; Lin et al., 2008). The working
electrode may also be modied with immobilised gold nanoparticle
for signal amplication (Lin et al., 2008). After binding with the
bacteria, secondary labelled antibodies are introduced. Alkaline

phosphatase labelled antibodies have been used as ampliers in
milk microbiology (Crowley et al., 1999; Luo et al., 2012) but horse
radish peroxidase (HRP) is the most commonly used enzyme. The
electrical current developed by the enzyme oxidation of a substrate,
such as hydrogen peroxide, is monitored and the intensity of the
signal is associated with bacteria concentration. Laczka et al. (2011)
published on a amperometric biosensor variant, where a 100 mL
diluted milk sample inoculated with E. coli was left to react with antiE. coli antibodies functionalised magnetic beads and with HRP
antibodies. The mixture was injected in a microuidic system where
magnets were located upstream from the electrode. The detection
limit was 100 cells/mL and the assay time was 1 h. This conguration
also achieved a good reproducibility, a common biosensor limit, with
only 1.3% relative standard deviation. With the same biorecognition
principle, Chemburu et al. (2005) published on a ow through
immunoassay where highly dispersed carbon particles provided a
large surface platform for a sandwich immunoassay. The functionalised carbon particles were placed on a disposable centrifuge lter.
Downstream the working electrode could detect the enzymatic
activity. The set-up was developed for the detection of three
pathogens: E. coli, L. monocytogenis and C. jejuni. The detection limits
in whole milk were of 50, 10 and 50 cells/mL and an assay time
of 30 min but the working range was limited to 600, 1500 and
500 cells/mL. The authors also found that the whole milk sample had
to be diluted 10-fold to prevent clogging of the lter membrane and
the deposition of particulates over the dispersed carbon layer. The
need for dilution is a drawback, as one of the driving factors for
pathogen analysis in samples of high volume is the preconcentration
requirement. Therefore, the use of lters is discouraged. However, it
is important to underline that the upstream separation of bacteria
and biorecognition element from the working electrode, developed
by Laczka et al. and Chemburu et al., provided an improvement in
Table 4
Selected research articles on the topic of biosensors in milk microbiology. Abbreviations: D.L.: detection limit; L.R.: linear range; A.T.: assay time; S.S.: storage stability; W.S.:
working stability.
Microorganism Sample type
and volume
Sample preparation
Electrochemical biosensors
E. coli
Milk
Diluted milk and
100 mL
immunomagnetic capture
E. coli
Semi-skimmed Filtrated milk and wash
milk
2 mL
Salmonella
spp.
S. aureus
Skimmed milk
5 mL
Raw milk
50 mL
Optical based biosensors

E. coli O157:
Fat free milk
H7
2 mL
E. coli O26
Pasteurised
skimmed milk
S. enteritidis
2 mL
Mass sensitive biosensors
E. coli O157:
Sterilised milk
H7
10 mL
S.
typhimurium
UHT fat free

milk
1 mL
8 h Enrichment and
immunomagnetic capture
None
None
None
Transducer
Biorecognition
element
Performance
factors
Reference
Amperometry
HRP antibodies
sandwich assay
Aptamer
D.L.: 100 cells/mL

A.T.: 1 h
D.L.: 6 CFU/mL
Laczka et al. (2011)
Potenziometry
Zelada-Guilln et al.
(2010)
Amperometry
Amperometry
HRP antibodies
sandwich assay
HRP antibodies
competitive assay
Cellphone camera with Antibodies sandwich

lens
with quantum dot
Surface plasmon
Antibody
resonance
L.R.: 610 CFU/mL

W.S.: 5 times
D.L.: 0.1 CFU/mL
A.T.: 50 min
D.L.: 1 CFU/mL
L.R.: 45 104 CFU/mL
A.T.: 2 h
D.L.: 10 CFU
A.T.: 2 h
D.L.: 2325 cells/mL
Libana et al. (2009)

Esteban-Fernndez de
vila et al. (2012)
Zhu et al. (2012)

Waswa et al. (2006)
A.T.: 1 h
Immunomagnetic
capture
Piezoelectric quartz
crystal microbalance
Antibody
None
Magnetoelastic sensor
Bacteriophage
D.L.: 53 CFU/mL
A.T.: 4 h
S.S.: 5 days at 4 1C
W.S.: 10 times
D.L.: 5 103 CFU/mL
A.T.: 20 min
L.R.: 5 1035 107 CFU/mL
Shen et al. (2011)
Lakshmanan et al.
(2007)
terms of sensor reproducibility, mass transfer to the electrode, higher

current and protection from fouling and passivation. A biosensor
based on anti-protein A antibodies proved its effectiveness in
selective detection of viable cells (Esteban-Fernndez de vila
et al., 2012). The biosensor was developed for the detection of
S. aureus in a 50 mL raw milk sample and showed a good detection
limit of 1 CFU/mL and poor sensitivity for lysed cells.
Potentiometry and aptamers were combined in a biosensor for
the detection of E. coli in semi-skimmed milk published by ZeladaGuilln et al. (2010). The aptamers were covalently linked to the
electrode surface via single-walled carbon nanotubes. 2 mL of
inoculated sample was used for the analysis after a brief sample
preparation that included removal of the milk matrix by ltration,
followed by a wash, then elution. An excellent detection limit of
6 CFU/mL was reported. The linear range extended to 104 CFU/mL.
The response time was of 2 min. Furthermore, the biosensor could be
reused 5 times. Regeneration was carried out with NaCl for 30 min.
3.2.2. Optical based biosensors
Optical based biosensors are commonly considered to be able to
achieve higher sensitivity than other biosensors but milk turbidity
and protein fouling are the limiting factors in their application for
milk testing.
BioFlashs is a commercially available optical biosensor by
Innovative Biosensor Environment Group Inc. Pathogen detection
is carried out with specic antibodies and calcium sensitive
bioluminescent molecules, which are detected using an easy to
use luminometer (Rider et al., 2003).
Surface plasmon resonance (SPR) circumvents the matrix
turbidity effect by measuring the refractive index on the reverse
side of the metal lm where the biological selective element is
immobilised (Mazumdar et al., 2007). SPR is commercially available and has been employed for the detection of foodborne
pathogens, namely BIACORE Q by BIACORE and Spreeta by
Sensata Technologies Inc. (Chinowsky et al., 2003). The reviewed
SPR biosensors used antibodies as biorecognition elements
(Waswa et al., 2007, 2006). Waswa et al. published on two label
free SPR biosensors for the detection of S. enteritidis and E. coli.
2 mL of articially contaminated skimmed milk was analysed
without any sample preparation. 23 and 25 cells/mL detection
limits were achieved. The response time was 1 h. Of signicance
was the work published by Rodriguez-Emmenegger et al. (2011),
where a modication utilising brushes on the sensing layer was
suggested to minimise milk fouling and the generated signal noise.
Cell phone technology was used with optical biosensors by Zhu
et al. (2012) in a portable and inexpensive set-up. The cell phone was
modied with an additional lens, ultraviolet LEDs and casing for
capillary assay. A disposable capillary worked as platform
for a quantum dot based sandwich immunoassay and as light
waveguide. The detection of E. coli O157:H7 was carried out in
2 h and the detection limit in fat free milk was 10 CFU/mL. This setup approaches the realisation of a portable and user friendly biosensor
the analysed volume was 2 mL, which may be enough for a
screening test, but new research toward sample preconcentration
and its integration within the system could complete the application.
Finally, the lab on a chip technology applied by Choi et al.
(2006) to raw milk achieved simultaneous detection of E. coli 0157:
H7 and Streptococcus agalactiae, antibiotic residues, somatic cells
and pH. The lab-on-chip incorporated microelectromechanical
systems, microuidics, immunouorescent reaction and uorescent microscopy.
3.2.3. Mass sensitive biosensors
Mass sensitive transducers such as the piezoelectric quartz
crystal microbalance (Chen et al., 2008; Shen et al., 2011) and
15
magnetoelastic transducers (Guntupalli et al., 2007; Lakshmanan

et al., 2007) have been applied to biosensors for milk microbiology
testing.
A piezoelectric biosensor was applied to the detection of E. coli
O157:H7 in a 1 mL milk sample (Shen et al., 2011). The authors
used an antibody functionalised crystal for the selective detection
of the pathogen. To enhance the detection, the coliform was rst
captured, concentrated and separated from the milk matrix in an
immunomagnetic reaction. The magnetic nanoparticles were also
loaded with biotin antibodies. Streptavidin gold was then conjugated to the labelled bacteria and the complex was treated
for gold catalytic growth. The detection limit was of 53 CFU/mL
and was achieved in 4 h of total assay time. A similar approach
was also adopted by Chen et al. (2008) for the detection of PCR
amplicons.
Magnetoelastic sensors are characterised by remote sensing as
the signal transduction is carried out from a distance with a coil
(Grimes et al., 2011). Magnetoelastic biosensors are the rst
example of wireless biosensing, which makes them remarkable
for biosensor platforms. Magnetoelastic biosensors have been
applied to S. typhimurium detection in fat free milk (Guntupalli
et al., 2007; Lakshmanan et al., 2007). In one study the sensor
surface was functionalised with antibodies while in the other the
selective layer was obtained with the immobilisation of lamentous bacteriophages. The biosensors had similar performance: the
detection limit was at 5 103 CFU/mL, the linear range extended
up to 5 107 CFU/mL and the detection time was 20 min for the
analysis of 1 mL sample. However, the bacteriophage biosensor
displayed a higher sensitivity, calculated as the slope of the linear
range. Furthermore, bacteriophage and phage peptides claim
higher stability, do not suffer from milk matrix inhibitory affect
and are cheaper (Smartt and Ripp, 2011).
3.3. Other sensing technologies
3.3.1. Flow cytometry
Flow cytometry is a fast technique and sensitive but its high
cost and signal to noise ratio in complex matrices have prevented
its uptake by the food industry.
On chip ow cytometry is commercially available and produced
by Agilent Technologies.
A selection of research papers on the topic of ow cytometry
for milk microbiological analysis is listed in Table 5.
Milk proteins and lipids are well known for binding staining
molecules and interfering with sample detection (Gunasekera,
2003; Yamaguchi et al., 2003). Thus milk samples must be prepared
for ow cytometry analysis. The most commonly adopted method
for sample preparation was developed by Gunasekera et al. (2000).
In this method, protein removal was achieved by applying an
enzymatic reaction for protein digestion with proteinase K and a
surfactant (Trixton X-100). The lipids were removed by centrifugation. Gunasekera et al. reported that the clearing method was
efcient on UHT milk but partially effective on raw milk. It was
suggested that the protein denaturation caused by the high temperature applied to the UHT milk facilitated the enzymatic digestion. Commercial milk clearing solutions have also been used (Holm
and Jespersen, 2003; McClelland and Pinder, 1994).
Various uorescent labelling methods have been applied. Dual
staining with uorescent dyes was applied to discriminate between
Gram positive and Gram negative bacteria (Gunasekera, 2003; Holm
and Jespersen, 2003). In the same publication, Gunasekera et al.
applied a second staining method based on uorescent in situ
hybridisation (FISH) using a 16S rRNA oligonucleotide probe specic
to eubacteria. Most importantly, the staining method could differentiate between viable and damaged cells. The analysis of 100 mL of
UHT milk was carried out in 1 h, the detection limit was 102 cells/mL
16
Table 5
Selected research articles on the topics of sensing ow cytometry, spectrometry techniques and multisensory systems in milk microbiology. Abbreviations: D.L.: detection
limit; L.R.: linear range; A.T.: assay time.
Microorganism
Flow cytometry
Gram /Gram
differentiation
Viable/damaged
eubacteria
differentiation
E. coli
Salmonella
Campylobacter
Listeria
E. coli O157:H7
Spectrometry techniques
S. typhimurium
C. jejuni
Multisensory systems
(Application to bacterial
spoilage)
(Application to clinical
mastitis detection)
Sample type and

volume
Sample preparation
Technology
Performance factors
Reference
Raw and UHT milk

100 mL
(I)
(II)
(III)
(IVa)
Proteinase K
Trixton X-100
Centrifuged and pelleted
Dual staining for viable/
dead cells differentiation
(IVb) Dual staining for Gram
7 differentiation
Dual staining and FISH based

ow cytometry
D.L.: 102 cells/mL

L:R.: 102106 cells/mL
A.T.: 1 h
Gunasekera
(2003)
Skimmed milk
90 mL
(I) Sample dilution

(II) Immunomagnetic separation
Multiplex antibodies functionalised

paramagnetic microsphere based
ow cytometry
D.L.: 105107 CFU/mL
Kim et al.
(2010)
UHT milk
100 mL
(I)
(II)
(III)
(IV)
Fluorescein labelled antibodies and

respiratory activity staining based
ow cytometry
D.L.: 103 cells/mL
Yamaguchi
et al. (2003)
Vitamin D fortied Dilution with selective

whole milk
medium broth
0.5 mL
Impedance microbiology with

interdigitated electrodes and
temperature control
A.T.: 10 h
Yang et al.
(2004)
2% Fat milk
1 mL
Spectrouorometer and quantum

dot functionalised aptamer
sandwich assay in adherent
cuvette
Proteinase K
Trixton X-100
Centrifuged and pelleted
Dual staining
(I) Dilution
(II) Magnetic capture
Skimmed milk
None
Raw milk
None
and the linear range comprised between 102 and 106 cells/mL.
A similar FISH based staining was applied to the detection of
L. monocytogenes with the Agilent Technologies microuidic system
(Ikeda et al., 2009). Results similar to the previous work were
conrmed. Differentiation between viable and dead cells was also
achieved by Yamaguchi et al. (2003) using staining for the detection
of respiratory activity of E. coli O157:H7 in 100 mL of UHT milk.
Immunomagnetic separation was applied in combination with a
multiplex sandwich immunoassay (Kim et al., 2010). The superparamagnetic beads were uorescently coded and the selective and
simultaneous identication of Salmonella, Campylobacter, E. coli and
Listeria was carried out in 90 mL of skimmed milk.
3.3.2. Impedance microbiology

Impedance microbiology is an action method for the specic
detection of bacterial growth. It is a sensitive and relatively rapid
method. Impedance microbiology was the rst action method for
Salmonella detection in food samples to be granted by the
association of Ofcial Analytical Chemists International (AOAC,
1995). Research in impedance microbiology has combined the
technique with the advancements offered by electrode microfabrication, microuidics, bio-functionalised magnetic capture and
dielectrophoresis for all electrode concentration methods. Recent
publications on the subject of impedance microbiology for foodborne pathogen detection have been reviewed by Yang and Bashir
(2008).
Electronic nose: array of 14

conducting polymer sensors
Electronic tongue : array of 15
potentiometric chemical sensors
D.L.: 1 CFU/mL
A.T.: 2 h
D.L.: 106 CFU/mL
D.L.: 10250 CFU/mL
A.T.: 1020 min
S.S.: long in
lyophilised form
D.L.: 103 CFU/mL
A.T.: 5 h
Sensitivity: 93%
Specicity: 96%
(103104 CFU/mL
inoculum)
Bruno et al.
(2009)
Magan et al.
(2001)
Mottram et al.
(2007)
Two papers were found for the detection of pathogens in milk.

An automated, portable and remote system for in-eld application
was developed by Grossi et al. (2011). The system included an
impedance measurement circuit board, thermoregulation, incubation chamber and microuidics. It was applied to the analysis of
4 mL raw milk for the detection of mesophilic and psychrotrophic
bacteria. No sample preparation was needed, although sample
dilution in growth medium decreased the assay time. 104 CFU/mL
mesophilic strains where detected in 10 h and 107 CFU/mL in 2 h.
The detection of psychrotrophic strains was longer: 30 and 4.5 h
for 104 and 107 CFU/mL respectively. Yang et al. (2004) published
on the detection of S. typhimurium in whole milk with an
interdigitated microelectrode impedance sensor. Selectivity was
achieved by sample dilution in selective medium. 0.5 mL samples
were analysed. 1 CFU/mL was detected in 10 h and 106 CFU/mL
in 2 h.
3.3.3. Spectrouorometry
A spectrouorometer was applied by Bruno et al. (2009) for the
detection of C. jejuni with DNA aptamers functionalised magnetic
beads and quantum dot functionalised aptamers in a sandwich
assay. To avoid the matrix affect, 1 mL of 2% fat milk sample was
diluted and mixed to react with the functionalised aptamers.
The sample was then introduced in an adherent plastic cuvette
and an external magnetic eld was applied. The reading was done
on a handheld spectrouorometer (Picouor) in 20 min. The
detection limit was between 10 and 250 CFU/mL in various food

matrices, including milk. The functionalised aptamers were stable
for long time in lyophilised form. The sensing technology is fast,
sensitive and stable. Cost effective and simple to use, it could be
applied to eld use as a screening method. Sample preconcentration is suggested to achieve good sample representability, detection probability and laboratory application.
3.3.4. Multisensory systems
Multisensors such as electronic nose for sensing gases and
electronic tongue for liquid samples, are coupled to chemometric
processing and have been applied to bacterial population analysis
in milk. The electronic nose and electronic tongue are secondary
methods, where the collected data are statistically analysed and
a signature is identied in comparison to a reference database.
Electronic sensors satisfy various requirements of the dairy industry, such as robustness, are hygienic, sterilisable and do not require
sample pretreatment. Reviews on electronic nose applications
written by Wilson and Baietto (2009) and electronic tongue in
food analysis written by Escuder-Gilabert and Peris (2010) will
offer further information on the technologies.
In the specic eld of foodborne pathogen detection in milk
(Table 4), Magan et al. (2001) published on an electronic nose
based on 14 conductive polymer sensors. The array was applied
to the analysis of skimmed milk inoculated with Pseudomonas
aureofaciens, P. fuorescens and Bacillus cereus for 5 h at 30 1C.
The application was successful in the differentiation between
unspoiled and spoiled milk and also between bacteria, when an
initial inoculation between 103 and 104 CFU/mL was used. Korel
et al. published on the application of the electronic nose for the
detection of odour development and quantication of microbial
loads down to 102 CFU/mL in milk (Korel and Balaban, 2002).
An electronic tongue was used by Mottram et al. (2007) for the
detection of clinical mastitis. The array was based on 15 potentiometric chemical sensors for inorganic and organic ions with
polyvinyl chloride membranes. The electronic tongue was able to
discriminate with sensitivity and specicity between milk from
healthy and mastitic glands. Wei et al. (2013) recently published
on the application of the electronic tongue for total bacterial count
and milk quality monitoring. The system could detect with good
linearity and prediction between 102 and 1010 CFU/mL.
4. Lab on chip
Microuidic analytical systems such as lab on chip (LOC)
include devices that integrate various operations on a small scale.
LOC performs laboratory functions such as sample handling,
mixing, reactions, separation and detection via their integration
and analyses parallelisation in a single device. Miniaturisation,
integration and automation have strong potential in terms of assay
time, thus high throughput and ease of use, sensitivity, portability,
laboratory based, eld and in-line application (Mairhofer et al.,
2009; Yoon and Kim, 2012).
All of the commercially available systems for pathogen detection listed in Table 1 are based on microuidic devices. In addition,
MiniFAB (www.minifab.com.au) is an enterprise that offers custom
development and manufacture of microuidic and microengineered products.
Various publications on LOC are included in this review work
on pathogen detection in milk. Essentially, most of the detection
methods have been integrated in LOC, including electrochemical
and optical detectors, as well as magneto-sensors.
Multiplex PCR amplicon detection from milk samples was
performed (Jin et al., 2009; Lee et al., 2008). Jin et al. published
on a multiplex bead-based mesouidic system where specic
17
oligonucleotide probes were immobilised for eight different

pathogens. The uorescent signal generated from the hybridisation reaction was obtained simultaneously from the multiple and
parallel microchannel system in 30 min. Also DNA extraction and
purication from whole milk samples, intended for PCR application, was carried out in a IFAST microuidic device (Strotman et al.,
2012). The technology exploited surface tension control over
gravity that takes place at micro-scale levels. The method proved
easy and fast.
Biosensors are particularly suitable for integration in LOC. In
Chemburu et al. (2005) a ow-through immunosensor was integrated with a peristaltic pump, reagent buffer vials, multi-way
valves and a potentiostat circuit. It was portable and automated.
The system proved sensitive and fast. In Laczka et al. (2011) a
biosensor was incorporated in a 3D microuidic system made
in PDMS.
Also, ow cytometry is a technology that by its nature can be
associated to LOC. Their combined application to L. monocytogenes
detection in milk has been reported by Ikeda et al. (2009). The
assay detection limit was 102 CFU/mL and was completed in 5 h,
including off chip milk clearing, hybridisation reaction and on chip
analyses.
The impedance microbiology system by Grossi et al. (2011) was
developed as a portable device based on LOC integrated impedance microbiology analyses. The method itself was lengthy but its
integration with a sample preconcentration method, for instance
based on functionalised and selective beads, could fully exploit the
advantages of the action method.
5. Summary and conclusions

The dairy industry is in need of advanced technologies for
pathogen detection methods to apply on-site. Recently, innovative
technologies that depart from classical methods have been available in the market. Most of the commercial systems are laboratory
based instruments. X-MARK by nanoRETE Inc. was found to be
suitable for eld analysis. All of the above are off-line instruments.
No commercial instruments are suitable for on-line or in-line
microbiological analysis and therefore commercial opportunities
are available.
The major unsolved challenges associated for the development
of a suitable method for early pathogen detection in the dairy
industry are the complex milk matrix affects and the detection
probability of applied methods. To address matrix interference
neutralisation, bacteria purication and concentration, further
research in sample preparation is essential. Sample concentration
is particularly important in terms of sample representability and
method sensitivity. The combination of the different approaches
reported in this review is possible and lend themselves to
integration and automation; these include centrifugation and
immunomagnetic separation, functionalised 3D structures e.g.
pillar or herringbone, acoustic wave, laminar ow focusing and
dielectrophoresis. Filtration is subject to clogging and it is discouraged. The development of novel methods is also encouraged.
Concentration methods are required to achieve bacteria purication and concentration from sample volumes in the order of tens
of mL down to hundreds of mL.
Practical Detection Limit should be included in the performance characterisation of a method sensitivity (Rossmanith et al.,
2006). Practical Detection Limit is calculated in relation to the
detection limit and taking into account the probability to have the
analyte represented within the sample volume. For instance, in
the method developed by Daly et al. (2002), a detection limit of
5 CFU/mL with a sample volume of 100 mL is associated with only
0.5 cells present in the sample. The probability of capturing 1 cell
18
is then 50%. The Practical Detection Limit reports the detection

limit associated with 100% probability of detecting the analyte. In
this example, the Practical Detection Limit is then 10 CFU/mL, thus
twice the (theoretical) detection limit. As argued earlier in regard
to sample volume representability, the sample volume should be
considered carefully during experimental design as well as in
result interpretation.
High sensitivity and selectivity have been achieved by nucleic
acid amplication methods and biosensors. Its specicity is up to
100% accordance when compared to reference methods, with no
room for interpretation subjectivity. However, it is a labour intense
and time consuming activity: although the time required to
perform a PCR is between 30 and 90 min, the whole process,
including sample preparation, DNA extraction and amplicon analysis, may take up to 2 working days. This is still faster than some
ofcial classical methods, but far from near real time. PCR is a
reliable but expensive method. Recent advancements in automation and multiplexing could amortise its costs: mPCR carries out
simultaneous multiple bacteria detection, leading to high results
throughput, overall time saving and labour effort. Considerable
progress has been made on standard protocols but the technology
is still lacking quality reagents, equipment and analysis methods,
which affect standardisation of the overall results between laboratories. A major drawback of PCR is the limit to discriminate
between viable and dead cells, which may lead to an overestimation and false positives. rtPCR and vPCR have been developed to
overcome such limits and promising results have been obtained.
Overall, PCR is still limited to the skilled environment of the
specialised molecular biology laboratory. However, thanks to the
recent advancement in the detection of viable cells, automation
and multiplexing, as well as the ISO standarisation work, PCR
could soon become a reality in the dairy industry laboratories for
ofine analysis. Likewise, LAMP and NASBA are promising and fast
developing technologies. LAMP amplication is highly selective
and fast. Amplicons result in higher concentration than in a
correspondent PCR reaction, thus shortening the analysis time
and increasing the sensitivity. The amplication methodology is of
recent development but the prompt application to the microbiological analysis of milk has proven its validity in the eld. NASBA
presents the advantage of viable cell detection. Both methods
operate in isothermal condition and their experimental set-up is
prone to system integration. Their automation and transfer to on
chip technology as well as method and reagent standardisation,
could provide LAMP and NASBA as valuable technologies in the
near future.
Biosensors have proven highly selective and sensitive. The
response time is exceptionally short, delivering analysis results
in near real time. They can provide long shelf life, portability,
automation and ease of use. Furthermore, biosensors can be
suitable to eld analyses and disposable. Disposable biosensors
are also useful to contain cross contamination between samples.
However, their extensive use in routine analysis will lead to
disposal issues. Furthermore, biosensors have a limited working
stability. The development of disposable biosensors for industrial
application should be oriented toward reusable and sterilisable
modules integrated with delimited disposable modules. The
employment of biodegradable polymers for their production is
also advised. Biosensors have been suggested to be directly
applicable to the analysis of complex matrices, such as milk, but
interference with antibodies and milk fouling on the sensing
surface have been reported as limiting factors. A common drawback to biosensors is the analysed samples volume. Sample
volumes that were analysed in the majority of the reviewed
articles were between tens and hundreds of mL, with only a few
cases where the sample volumes were up to 2 mL. That suggests
poor sample representability and specically low probability to
capture the target pathogen within the sample, especially when

zero tolerance applies and a 25 g sample size is a legislative
requirement. As a result, biosensors suffer in detection probability,
which negatively impacts the overall method sensitivity and its
Practical Detection Limit. The performance parameters concerning
the method sensitivity should include detection limit, sensitivity,
detection probability and Practical Detection Limit as explained
above. Sample preconcentration methods should be applied in
preparative steps with the additional benet of limiting the matrix
affect. A nal issue that has limited biosensors application to
routine analysis is their poor reproducibility. Surface functionalisation with biorecognition elements has proven difcult to replicate with acceptable standard deviation. Encouraging results have
been obtained with competitive designs and beads functionalisation. Biosensors have high potential for eld analysis as well as
laboratory applications. The combination of fast response time,
low cost, ease of use, automation and integration in microuidic
systems, puts biosensors in a competitive place to become a useful
tool for pathogen early detection and to efciently process the
large sample quantities produced by the enforcement of HACCP.
Biosensors are suggested to be inexpensive thus relevant for
industrial applications, where a high number of samples need to
be analysed. Small and medium enterprises cannot afford the
management of specialised laboratories. Cost effective analytical
solutions such as biosensors could support them to full legislator
and safety requirements in an independent and time effective
manner.
Flow cytometry has some major advantages, such as the ability
to discriminate between viable and damaged or dead cells. It
provides multiplex analysis, precise data acquisition and reproducibility, short analysis times, automation and integration. Sample
preparation is required. The samples analysed in the reviewed
articles were of small volumes, e.g. 100 mL, and the reported
detection limits were high, e.g. 102 CFU/mL. Furthermore, ow
cytometry presents major limitations such as the signal to noise
ratio generated by complex matrices and high costs. However, ow
cytometry integration with sample preparation in a LOC setup
with parallel analysis managed within a microuidic system, could
move the technology toward laboratory application for screening
cell counts.
Impedance microbiology offers a combination of advantages.
First of all it is distinguished as an action method and allows the
measurement of bacteria growth activity along with its capability
to detect viable cells only. It combines selectivity, sensitivity,
automation, portability, remote sensing, potential for on-line
analysis and multiplex sensing with ease of use and low costs.
Impedance microbiology employs bare electrodes that are compatible with sterilisation methods and reusability, thus it is suitable
for repeated and automated analysis. Recent micro- and nanotechnology advancements in sample preparation and concentration have not yet been transferred to pathogen detection in milk
and this is encouraged. Impedance microbiology integration with
micro- and nanotechnologies, sample preparation and parallel
analysis in a LOC device could provide a multiplex action technology suitable for off-line and automated on-line analyses.
Of relevance is the handheld sensing technology developed by
Bruno et al. (2009). The technology proved fast, sensitive and with
long lasting storage stability. Cost effective and simple to use, it
could be applied to eld use as a screening method. Sample
preconcentration is suggested to achieve good sample representability, detection probability and off-line laboratory application.
Electronic nose and electronic tongue are secondary methods
that have poor sensitivity and selectivity but can withstand
prolonged stable operation with no maintenance, thus could be
suitable to in-line analyses for the monitoring of total bacteria
count, hygienic condition and milking mastitic animals. Electronic
A possible roadmap for the development of highly integrated

microsystems for microorganism detection based on compact
sample preparation, genetic material amplication and biosensor
detection is outlined here (Fig. 3). The roadmap was developed as
a result of this review work and taking into consideration the
FoodMicroSystems ndings that are discussed in the dairy industry report (FoodMicroSystems, 2013). The FoodMicroSystems roadmapping methodology was based on the NEXUS approach,
application driven and combined the state of the art with the
current research and developement. The main driving factors for
the roadmap suggested here are the potential application combined with system sensitivity and response time, forcasted to
achieve in-line analysis with less than 1 CFU/mL detection limit in
less than an hour by 2030. The current technologies and methods
are moving from the state of the art toward integrated bench-top
system for near-line analysis operated by trained personnel. This
development step requires the integration of sample preparation
with on chip DNA/RNA amplication and detection. The next level
of development should provide a handheld system suitable for online analysis. The system should be easy to operate by trained staff
and the costs should decrease considerably. In order to achieve
such a progression, the sample preparation module should be
compact, the genetic material extraction and amplication should
be based on immobilised enzymes and the detection should be
label free and based on biosensors. Finally, automatic in-line
systems, perhaps operated remotely, should be highly cost effective. The system integration should be of high level and the system
becomes part of the production management and chain. The rst
step of the roadmap is based on work in progress while the
following stages are idealistic and anticipate the developmet of
likely systems. Other technology development streams will be
possible and expectable but always application driven. On-line and
in-line systems could nd application at different levels in the
production chain. These innovative systems could be installed as
part of milking robot to monitor mastitic animal. System set-ups in
refrigerated storage tanks could provide quality control at dairy
sites and from milk vending machines. A handheld system could
be used by drivers of milk vans and should be able to carry out the
analysis during loading or transport. Fast systems could be applied
at reception and delivery points of dairy processing. In that case,
the system should be cartridge based, with a maximum cost of 5
(FoodMicroSystems, 2013).
It was the intention of this review work to stimulate creative
thinking in the research community and to extend the outstanding
sensors hold the promise of a reagentless future system, which

could be remotely managed, distributed over the industry at
HACCP, integrated with the production management and that
could provide a constant stream of in-line real time results.
In conclusion, LOC holds the potential to develop sensitive, fast,
cost effective, user-friendly and portable analytical systems,
applicable either to laboratory analyses, eld tests and on-line
analysis. The integration of hybrid technologies, such as compact
sample preparation methods with miniaturised detection technologies via meso and microuidic pathways can decrease processing
time and sample movement. Sensing system integration and
automation have also emerged as key points in this review work,
along with multiplexing and sample preparation. In that respect,
LOC holds the ultimate promise to full requirements of regulators
and industry. Unsolved challenges associated to LOC are reported.
In particular, small particles, precipitates and bubbles can easily
block microuidic systems. The integration of sample preparation
is a crucial step to provide a fully automated and compact system
but still remains an unsolved challenge.
6. Future perspectives
Starting from the methods and technologies that are commonly
used in laboratory, development paths toward the next generation
of analytical systems for pathogens detection in milk will aim to
(i) produce compact and automated modules for sample preparation module, (ii) reduce the total analysis response time and
(iii) provide direct detection with the automated and integrated
system. Such technology could stimulate various applications at
off-line, on-line and even in-line levels. A feasable application in
the near future could be as screening tests in conjunction with the
existing methods.
Dispite the limited number of micro-monolytical systems for food
safety existing in the market, there is an increasing demand for
innovation. In the coming years, more analytical systems will be
commercially available. Interesting concepts will be developed to
satisfy the variety of market demand. It is reasonable to expect the
development of common platform with exchangeable and/or disposable small components, such as sensing units and cartridges,
taylored for multiparametric tests and applications. The combination
of nanomicrotechnologies for sensing, uidics and biosurfaces with
ICT will provide further advancements (FoodMicroSystems, 2013).
Compact sample
prep bioreactor
Multiplexing label free biosensing
System integration:
On-chip integration
- Sample preparation
Miniaturisation
System integration:
- HACCP and production
- Sample prep on-chip
<1h
Automation
Time to result
Sensitivity
<1 CFU/ml
- On-chip DNA/RNA
extraction
- Amplification
100 CFU/ml
2000
19
5 days
2010
2020
2030
Years
Fig. 3. Roadmap of highly integrated heterogeneous microsystems for microorganisms detection based on compact sample preparation, genetic material amplication and
biosensor detection.
20
achievements obtained so far. New and advanced sensing systems

for pathogen detection in milk and the dairy industry could full
the industry need for a cheaper and faster detection technology.
The technology could nd application in precision dairy farming
and production management. Affordable and effective quality
control could also promote the development of sustainable small
and medium production plants, for instance by providing health
conscious consumers with certied raw milk.
Acknowledgement
The authors sincerely acknowledge the Fondazione Bruno
Kessler and the European Commission who have supported Alessia
Mortari and her project IM-Milk with the cofund Marie Curie
Actions RESTATE project (Grant agreement no. 267224).
References
Ackers, M.L., Schoenfeld, S., Markman, J., Smith, M.G., Nicholson, M.A., DeWitt, W.,
Cameron, D.N., Grifn, P.M., Slutsker, L., 2000. J. Infect. Dis. 181, 18341837.
Alocilja, E.C., Radke, S.M., 2003. Biosens. Bioelectron. 18, 841846.
Amagliani, G., Brandi, G., Omiccioli, E., Casiere, A., Bruce, I.J., Magnani, M., 2004.
Food Microbiol. 21, 597603.
Amagliani, G., Omiccioli, E., Campo, A., Bruce, I.J., Brandi, G., Magnani, M., 2006.
J. Appl. Microbiol. 100, 375383.
AOAC, 1995. Salmonella in food, automated conductance methods. Method 991.38.
In: Ofcial Methods of Analysis of AOAC International. Association of Ofcial
Analytical Chemists International, Gaithersburg, pp. 9294.
Boyer, M., Combrisson, J., 2013. Int. Dairy J. 30, 4552.
Bruno, J.G., Phillips, T., Carrillo, M.P., Crowell, R., 2009. J. Fluoresc. 19, 427435.
Chemburu, S., Wilkins, E., Abdel-Hamid, I., 2005. Biosens. Bioelectron. 21, 491499.
Chen, S.-H., Wu, V.C.H., Chuang, Y.-C., Lin, C.-S., 2008. J. Microbiol. Methods 73,
717.
Chinowsky, T.M., Quinn, J.G., Bartholomew, D.U., Kaiser, R., Elkind, J.L., 2003. Sensor
Actuators B: Chem. 91, 266274.
Choi, J.-W., Kim, Y.-K., Kim, H.-J., Lee, W., Seong, G.H., 2006. J. Microbiol. Biotechnol.
16, 12291235.
Choi, S.H., Lee, S.B., 2004. J. Anim. Sci. Technol. 46, 841848.
Compton, J., 1991. Nature 350, 9192.
Cremonesi, P., Castiglioni, B., Malferrari, G., Biunno, I., Vimercati, C., Moroni, P.,
Morandi, S., Luzzana, M., 2006. J. Dairy Sci. 89, 163169.
Crowley, E.L., O'Sullivan, C.K., Guilbault, G.G., 1999. Analyst 124, 295299.
Daly, P., Collier, T., Doyle, S., 2002. Lett. Appl. Microbiol. 34, 222226.
De Buyser, M.-L., Dufour, B., Maire, M., Lafarge, V., 2001. Int. J. Food Microbiol. 67,
117.
Debarry, J., Garn, H., Hanuszkiewicz, A., Dickgreber, N., Blmer, N., von Mutius, E.,
Bufe, A., Gatermann, S., Renz, H., Holst, O., Heine, H., 2007. J. Allergy Clin.
Immunol. 119, 15141521.
Dineen, S.S., Takeuchi, K., Soudah, J.E., Boor, K.J., 1998. J. Food Prot. 61, 16021608.
Djnne, B., Jensen, M.R., Grant, I.R., Holstad, G., 2003. Vet. Microbiol. 92, 135143.
Duarte, C., Salm, E., Dorvel, B., Reddy, B., Bashir, R., 2013. Biomed. Microdevices 15,
821830.
EC No. 853/2004. Laying down specic hygiene rules for on the hygiene of
foodstuffs. Ofcial Journal of the European Union.
EC No. 2073/2005. Microbiological criteria for foodstuffs. Ofcial Journal of the
European Union.
EC Agriculture and Rural Development, 2005. Organic Farming in the European
Union Facts and Figures.
EC Health & Consumer Protection General Directorate, 2005. Guidance Document
Implementation of Procedures Based on the HACCP Principles, and Facilitation
of the Implementation of the HACCP Principles in Certain Food Businesses.
Ege, M.J., Mayer, M., Schwaiger, K., Mattes, J., Pershagen, G., van Hage, M.,
Scheynius, A., Bauer, J., von Mutius, E., 2012. Allergy 67, 15651571.
Ehrs, S., gren, P., Stephansson, O., Garbom, S., Zumpe, A.L., 2011. J. Food Saf. 31,
225231.
Escamilla-Gmez, V., Campuzano, S., Pedrero, M., Pingarrn, J., 2008. Talanta 77,
876881.
Escuder-Gilabert, L., Peris, M., 2010. Anal. Chim. Acta 665, 1525.
Esteban-Fernndez de vila, B., Pedrero, M., Campuzano, S., Escamilla-Gmez, V.,
Pingarrn, J.M., 2012. Anal. Bioanal. Chem. 403, 917925.
Falentin, H., Postollec, F., Parayre, S., Henaff, N., Le Bivic, P., Richoux, R., Thierry, A.,
Sohier, D., 2010. Int. J. Food Microbiol. 144, 1019.
Farm-to-Consumer Legal Defence Fund, 2013. Raw Milk Nation http://farmtocon
sumer.org/raw_milk_map.htm (accessed 06.03.14.).
FDA, 1997. Dairy Grade a Voluntary HACCP http://www.fda.gov/Food/GuidanceR
egulation/HACCP/ucm2007982.htm (accessed 06.03.14.).
FDA, 2012. Bad Bug Book, Foodborne Pathogenic Microorganisms and Natural
Toxins, second ed. http://www.fda.gov/Food/FoodborneIllnessContaminants/
CausesOfIllnessBadBugBook/ (accessed 06.03.14.).
FoodMicroSystems, 2013. Road Map Section 1 Implementation of Mycrosystems in

the Dairy Sectos http://www.foodmicrosystems.eu/?page_id=3063 (accessed
06.03.14.).
Food Standards Australia New Zealand, 2009. Proposal P1007 Primary Production
and Processing Requirements for Raw Milk Products Assessment of Potential
Health Benets Associated with Raw Milk.
Food Standards Australia New Zealand, 2012. Approval Report Proposal P1007
Primary Production & Processing Requirements for Raw Milk Products.
Fox, P.F., 2011. In: Fuquay, J., Fox, P., McSweeney, P. (Eds.), Encyclopedia of Dairy
Sciences, second ed. Academic Press, San Diego, pp. 458466.
Garca-Nogales, P., Serrano, A., Secchi, S., Gutirrez, S., Ars, A., 2010. Electron.
J. Biotechnol. 13, 17.
Giannino, M.L., Aliprandi, M., Feligini, M., Vanoni, L., Brasca, M., Fracchetti, F., 2009.
J. Microbiol. Methods 78, 181188.
Grossi, M., Lanzoni, M., Pompei, A., Lazzarini, R., Matteuzzi, D., Ricc, B., 2011. 4th IEEE
International Workshop on Advances in Sensors and Interfaces. pp. 132137.
Grimes, C.A., Roy, S.C., Rani, S., Cai, Q., 2011. Sensors 11, 28092844.
Gunasekera, T., 2003. Int. J. Food Microbiol. 85, 269279.
Gunasekera, T.S., Atteld, P.V, Veal, D.A., 2000. Appl. Environ. Microbiol. 66,
12281232.
Guntupalli, R., Lakshmanan, R.S., Johnson, M.L., Hu, J., Huang, T.-S., Barbaree, J.M.,
Vodyanoy, V.J., Chin, B.A., 2007. Sens. Instrum. Food Qual. Saf. 1, 310.
Hein, I., Flekna, G., Krassnig, M., Wagner, M., 2006. J. Microbiol. Methods 66,
538547.
Holm, C., Jespersen, L., 2003. Appl. Environ. Microbiol. 69, 28572863.
Hsih, H.-Y., Tsen, H.-Y., 2001. J. Food Prot. 64, 17441750.
Ikeda, M., Yamaguchi, N., Nasu, M., 2009. J. Health Sci. 55, 851856.
ISO, 2005a. ISO 22174:2005 Microbiology of Food and Animal Feeding Stuffs
Polymerase Chain Reaction (PCR) for the Detection of Food-Borne Pathogens
General Requirements and Denitions.
ISO, 2005b. ISO/TS 20836:2005 Microbiology of Food and Animal Feeding Stuffs
Performance Testing for Thermal Cyclers.
ISO, 2006a. ISO 20837:2006 Microbiology of Food and Animal Feeding Stuffs
Requirements for Sample Preparation for Qualitative Detection.
ISO, 2006b. ISO 20838:2006 Microbiology of Food and Animal Feeding Stuffs
Requirements for Amplication and Detection for Qualitative Methods.
ISO, 2008. ISO 707:2008 Milk and Milk Products Guidance on Sampling.
ISO, 2011. ISO 22119:2011 Microbiology of Food and Animal Feeding Stuffs RealTime Polymerase Chain Reaction (PCR) for the Detection of Food-Borne
Pathogens General Requirements and Denitions.
Jin, S.-Q., Yin, B.-C., Ye, B.-C., 2009. Appl. Environ. Microbiol. 75, 66476654.
Khare, S., Ficht, T.A., Santos, R.L., Ficht, A.R., Zhang, S., Grant, I.R., Libal, M., Hunter,
D., Adams, L.G., Romano, J., 2004. J. Clin. Microbiol. 42, 10751081.
Kim, J.S., Taitt, C.R., Ligler, F.S., Anderson, G.P., 2010. Sens. Instrum. Food Qual. Saf. 4,
7381.
Korel, F., Balaban, M., 2002. J. Food Sci. 67, 758764.
Laczka, O., Maesa, J.-M., Godino, N., del Campo, J., Fougt-Hansen, M., Kutter, J.P.,
Snakenborg, D., Muoz-Pascual, F.-X., Baldrich, E., 2011. Biosens. Bioelectron.
26, 36333640.
Lakshmanan, R.S., Guntupalli, R., Hu, J., Petrenko, V. a., Barbaree, J.M., Chin, B.A.,
2007. Sensor Actuators B: Chem. 126, 544550.
Lazcka, O., Del Campo, F.J., Muoz, F.X., 2007. Biosens. Bioelectron. 22, 12051217.
Lee, K.-H., Lee, J.-W., Wang, S.-W., Liu, L.-Y., Lee, M.-F., Chuang, S.-T., Shy, Y.-M.,
Chang, C.-L., Wu, M.-C., Chi, C.-H., 2008. J. Vet. Diagn. Investig. 20, 463471.
LeJeune, J.T., Rajala-Schultz, P.J., 2009. Clin. Infect. Dis. 48, 93100.
Li, X., Zhang, S., Zhang, H., Zhang, L., Tao, H., Yu, J., Zheng, W., Liu, C., L, D., Xiang, R.,
Liu, Y., 2009. Int. J. Food Microbiol. 133, 252258.
Li, Y., Jiang, M., Liu, W., Zhang, L., Zhang, S., Zhao, X., Xiang, R., Liu, Y., 2010. Mol. Cell.
Probe 24, 6871.
Libana, S., Lermo, A., Campoy, S., Corts, M.P., Alegret, S., Pividori, M.I., 2009.
Biosens. Bioelectron. 25, 510513.
Lin, Y.-H., Chen, S.-H., Chuang, Y.-C., Lu, Y.-C., Shen, T.Y., Chang, C.A., Lin, C.-S., 2008.
Liu, C., Zheng, W., Zhang, H., Hou, Y., Liu, Y.I.N., 2007. J. Food Saf. 29, 8394.
Luo, C., Lei, Y., Yan, L., Yu, T., Li, Q., Zhang, D., Ding, S., Ju, H., 2012. Electroanal 24,
11861191.
Magan, N., Pavlou, A., Chrysanthakis, I., 2001. Sensor Actuators B: Chem. 72, 2834.
Mairhofer, J., Roppert, K., Ertl, P., 2009. Sensors 9, 48044823.
Mandal, P.K., Biswas, A.K., Choi, K., Pal, U.K., 2011. Am. J. Food Technol. 6, 87102.
Mazumdar, S.D., Hartmann, M., Kmpfer, P., Keusgen, M., 2007. Biosens. Bioelectron. 22, 20402046.
McClelland, R.G., Pinder, A.C., 1994. Appl. Environ. Microbiol. 60, 42554262.
Morgan, D., Newman, C.P., Hutchinson, D.N., Walker, A.M., Rowe, B., Majid, F., 2009.
Epidemiol. Infect. 111, 181188.
Mottram, T., Rudnitskaya, A., Legin, A., Fitzpatrick, J.L., Eckersall, P.D., 2007. Biosens.
Bioelectron. 22, 26892693.
Mullis, K.B., Erlich, H.A., Arnheim, N., Horn, G.T., Saiki, R.K., Scharf, S.J., 1987. Patent
No. US 4,683,195.
Nam, H.-M., Srinivasan, V., Gillespie, B.E., Murinda, S.E., Oliver, S.P., 2005. Int. J. Food
Microbiol. 102, 161171.
Notomi, T., Okayama, H., Masubuchi, H., Yonekawa, T., Watanabe, K., Amino, N.,
Hase, T., 2000. Nucleic Acids Res. 28 (e63), ivii.
Notomi, T., Hase, T., 2002. Patent No. US 6,410,278 B1.
O'Grady, J., Lacey, K., Glynn, B., Smith, T.J., Barry, T., Maher, M., 2009. FEMS
Microbiol. Lett. 301, 218223.
Oliver, J.D., 2005. FEMS Microbiol. Rev. 34, 415425.
Omiccioli, E., Amagliani, G., Brandi, G., Magnani, M., 2009. Food Microbiol. 26,
615622.
Ongol, M.P., Tanaka, M., Sone, T., Asano, K., 2009. Food Res. Int. 42, 893898.
Paul, M., Van Hekken, D.L., Brewster, J.D., 2013. Int. Dairy J. 32, 5360.
Perelle, S., Dilasser, F., Malorny, B., Grout, J., Hoorfar, J., Fach, P., 2004. Mol. Cell.
Probe 18, 409420.
Prez-Lpez, B., Merkoi, A., 2011. Trends Food Sci. Technol. 22, 625639.
Postollec, F., Falentin, H., Pavan, S., Combrisson, J., Sohier, D., 2011. Food Microbiol.
28, 848861.
Quigley, L., O'Sullivan, O., Stanton, C., Beresford, T.P., Ross, R.P., Fitzgerald, G.F.,
Cotter, P.D., 2013. FEMS Microbiol. Rev. 37, 664698.
Rider, T.H., Petrovick, M.S., Nargi, F.E., Harper, J.D., Schwoebel, E.D., Mathews, R.H.,
Blanchard, D.J., Bortolin, L.T., Young, A.M., Chen, J., Hollis, M.A., 2003. Science
301, 213215.
Riyaz-Ul-Hassan, S., Verma, V., Qazi, G.N., 2013. Lett. Appl. Microbiol. 56, 275282.
Rodriguez-Emmenegger, C., Avramenko, O.A., Brynda, E., Skvor, J., Alles, A.B., 2011.
Rossen, L., Nrskov, P., Holmstrm, K., Rasmussen, O.F., 1992. Int. J. Food Microbiol.
17, 3745.
Rossmanith, P., Krassnig, M., Wagner, M., Hein, I., 2006. Res. Microbiol. 157,
763771.
Schmid, D., Fretz, R., Winter, P., Mann, M., Ho
ger, G., Sto
ger, A., Ruppitsch, W.,
Ladsttter, J., Mayer, N., de Martin, A., Allerberger, F., 2009. Wien. Klin.
Wochenschr. 121, 125131.
Shen, Z.-Q., Wang, J.-F., Qiu, Z.-G., Jin, M., Wang, X.-W., Chen, Z.-L., Li, J.-W., Cao, F.-H.,
2011. Biosens. Bioelectron. 26, 33763381.
Sheu, S.-J., Hwang, W.-Z., Chen, H.-C., Chiang, Y.-C., Tsen, H.-Y., 2009. J. Food Prot. 72,
93100.
21
Sheu, S.-J., Hwang, W.-Z., Chiang, Y.-C., Lin, W.-H., Chen, H.-C., Tsen, H.-Y., 2010.
J. Food Sci. 75, M521M527.
Smartt, A.E., Ripp, S., 2011. Anal. Bioanal. Chem. 400, 9911007.
Stevens, K.A., Jaykus, L.-A., 2004. Crit. Rev. Microbiol. 30, 724.
Strotman, L.N., Lin, G., Berry, S.M., Johnson, E.A., Beebe, D.J., 2012. Analyst 137,
40234028.
Ulve, V.M., Monnet, C., Valence, F., Fauquant, J., Falentin, H., Lortal, S., 2008. J. Appl.
Microbiol. 105, 13271333.
Upton, P., Coia, J., 1994. Lancet 344, 1015.
Vaitilingom, M., Gendre, F., Brignon, P., 1998. Appl. Environ. Microbiol. 64,
11571160.
Vanegas, M.C., Vsquez, E., Martinez, A.J., Rueda, A.M., 2009. Food Control 20,
430432.
Velusamy, V., Arshak, K., Korostynska, O., Oliwa, K., Adley, C., 2010. Biotechnol. Adv.
28, 232254.
Wang, D., Huo, G., Ren, D., Li, Y., 2010. J. Food Saf. 30, 251262.
Waswa, J., Irudayaraj, J., DebRoy, C., 2007. LWT Food Sci. Technol. 40, 187192.
Waswa, J.W., Debroy, C., Irudayaraj, J., 2006. J. Food Process. Eng. 29, 373385.
Wei, Z., Wang, J., Zhang, X., 2013. Electrochim. Acta 88, 231239.
Wilson, A.D., Baietto, M., 2009. Sensors 9, 50995148.
Yamaguchi, N., Sasada, M., Yamanaka, M., Nasu, M., 2003. Cytom. Part A 54, 2735.
Yang, C., Jiang, Y., Huang, K., Zhu, C., Yin, Y., 2003. FEMS Immunol. Med. Microbiol.
38, 265271.
Yang, L., Bashir, R., 2008. Biotechnol. Adv. 26, 135150.
Yang, L., Li, Y., Grifs, C.L., Johnson, M.G., 2004. Biosens. Bioelectron. 19, 11391147.
Yoon, J.-Y., Kim, B., 2012. Sensors 12, 1071310741.
Zelada-Guilln, G.A., Bhosale, S.V., Riu, J., Rius, F.X., 2010. Anal. Chem. 82,
92549260.
Zhu, H., Sikora, U., Ozcan, A., 2012. Analyst 137, 25412544.

Recent Sensing Technologies For Pathogen Detection in Milk A Review

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Recent Sensing Technologies For Pathogen Detection in Milk A Review

Uploaded by

Copyright:

Available Formats

Biosensors and Bioelectronics 60 (2014) 821

Contents lists available at ScienceDirect

Biosensors and Bioelectronics

Recent sensing technologies for pathogen detection in milk: A review

A. Mortari, L. Lorenzelli / Biosensors and Bioelectronics 60 (2014) 821

(EC Health & Consumer Protection General Directorate, 2005; FDA,

development for advanced and integrated analytical systems that

2. Milk sampling and sample preparation

A. Mortari, L. Lorenzelli / Biosensors and Bioelectronics 60 (2014) 821

Analytical reference method

Analysis time, days

7500 PCR System

Fig. 2. Publication percentiles on sensing methods for the detection of foodborne

relation to each technical section of this review. Yet, none of the

3. Methods for foodborne pathogen detection

A. Mortari, L. Lorenzelli / Biosensors and Bioelectronics 60 (2014) 821

Sample type and

End point PCR

Real Time Quantitative PCR

Sample preparation and

(I) Matrix separation by centrifugation

Amagliani et al. (2006,

(I) Matrix separation by centrifugation

D.L.: 0.1 CFU/mL

Djnne et al. (2003)

(I) Matrix separation by centrifugation

D.L.: 6-15 CFU/PCR tube

Yang et al. (2003)

Paul et al. (2013)

Khare et al. (2004)

Hein et al. (2006)

(I) Sample preparation using EDTA and

Matrix separation by centrifugation

(I) Enrichment culture

D.L.: 100 CFU/mL

Giannino et al. (2009)

hipO and mapA

(I) Enrichment culture

Omiccioli et al. (2009)

D.L.: 186 CFU/mL

Wang et al. (2010)

Sample preparation using EDTA, detergent

O'Grady et al. (2009a)

developed in this century for the detection of pathogens in milk.

The authors should point out that in many of the reviewed

A. Mortari, L. Lorenzelli / Biosensors and Bioelectronics 60 (2014) 821

3.1.1. Sample preparation

3.1.2. DNA extraction

methods (Amagliani et al., 2006, 2004; Djnne et al., 2003) is also

3.1.3. PCR variants and amplicons analysis

A. Mortari, L. Lorenzelli / Biosensors and Bioelectronics 60 (2014) 821

PCR is highly selective and is able to detect subdominant

rtPCR represents a step forward in the detection of viable cells.

A. Mortari, L. Lorenzelli / Biosensors and Bioelectronics 60 (2014) 821

purication. The most common transducers in milk analysis are

3.2.1. Electrochemical biosensors

bacteria, secondary labelled antibodies are introduced. Alkaline

Optical based biosensors

UHT fat free

D.L.: 100 cells/mL

Laczka et al. (2011)

Cellphone camera with Antibodies sandwich

L.R.: 610 CFU/mL