Recovery and detection of Cryptosporidium parvum oocysts from water samples using continuous flow centrifugation.

Abstract

Continuous flow centrifugation (CFC) was used in conjunction with immunomagnetic separation (IMS) and immunofluorescence microscopy (IFA) and nested PCR to recover and detect oocysts of Cryptosporidium parvum and cysts of Giardia intestinalis from 10L volumes of source water samples. Using a spiking dose of 100 oocysts, nine of 10 runs were positive by IFA, with a mean recovery of 4.4+/-2.27 oocysts; when another 10 runs were analyzed using nested PCR to the TRAP C-1 and Cp41 genes, nine of 10 were positive with both PCR assays. When the spiking dose was reduced to 10 oocysts in 10L, 10 of 12 runs were positive by IFA, with a mean oocyst recovery of 3.25+/-3.25 oocysts. When 10 cysts of Giardia intestinalis were co-spiked with oocysts into 10L of source water, five of seven runs were positive, with a mean cyst recovery of x=0.85+/-0.7. When 10 oocysts (enumerated using a fluorescence activated cell sorter) were spiked into 10L volumes of tap water, one of 10 runs was positive, with one oocyst detected. For the majority of the source water samples, turbidities of the source water samples ranged from 1.1 to 22 NTU, but exceeded 100 NTU for some samples collected when sediment was disturbed. The turbidities of pellets recovered using CFC and resuspended in 10 mL of water were very high (exceeding 500 NTU for the source water-derived pellets and 100 NTU for the tap water-derived pellets). While not as efficient as existing capsule-filtration based methods (i.e., US EPA methods 1622/1623), CFC and IMS may provide a more rapid and economical alternative for isolation of C. parvum oocysts from highly turbid water samples containing small quantities of oocysts.

Full text

Page 1

Water Research 37 (2003) 3551–3560
Recovery and detection of Cryptosporidium parvum oocysts
from water samples using continuous flow centrifugation
James A. Higgins*, James M. Trout, Ronald Fayer,
Daniel Shelton, Mark C. Jenkins
USDA-ARS, Rm 202, Bldg 173, 10300 Baltimore Blvd, Beltsville, MD 20705, USA
Received 1 December 2001; received in revised form 1 March 2003
Abstract
Continuous flow centrifugation (CFC) was used in conjunction with immunomagnetic separation (IMS) and
immunofluorescence microscopy (IFA) and nested PCR to recover and detect oocysts of Cryptosporidium parvum and
cysts of Giardia intestinalis from 10L volumes of source water samples. Using a spiking dose of 100 oocysts, nine of 10
runs were positive by IFA, with a mean recovery of 4.472.27 oocysts; when another 10 runs were analyzed using nested
PCR to the TRAP C-1 and Cp41 genes, nine of 10 were positive with both PCR assays. When the spiking dose was
reduced to 10 oocysts in 10L, 10 of 12 runs were positive by IFA, with a mean oocyst recovery of 3.2573.25 oocysts.
When 10 cysts of Giardia intestinalis were co-spiked with oocysts into 10L of source water, five of seven runs were
positive, with a mean cyst recovery of x ¼ 0:8570:7: When 10 oocysts (enumerated using a fluorescence activated cell
sorter) were spiked into 10L volumes of tap water, one of 10 runs was positive, with one oocyst detected. For the
majority of the source water samples, turbidities of the source water samples ranged from 1.1 to 22NTU, but exceeded
100NTU for some samples collected when sediment was disturbed. The turbidities of pellets recovered using CFC and
resuspended in 10mL of water were very high (exceeding 500NTU for the source water-derived pellets and 100NTU
for the tap water-derived pellets). While not as efficient as existing capsule-filtration based methods (i.e., US EPA
methods 1622/1623), CFC and IMS may provide a more rapid and economical alternative for isolation of C. parvum
oocysts from highly turbid water samples containing small quantities of oocysts.
Published by Elsevier Science Ltd.
Keywords: Continuous flow centrifugation; Cryptosporidium parvum; Detection
1. Introduction
A variety of techniques have been evaluated for the
recovery and detection of Cryptosporidium parvum
oocysts from source and finished water samples [1].
The most widely used technique has been cartridge
filtration in conjunction with immunomagnetic separa-
tion and immunofluorescence microscopy (i.e., US EPA
method 1622/23). An extensive evaluation of the US
EPA method 1622 for C. parvum detection was
conducted by two laboratories. One laboratory obtained
oocyst recoveries ranging from 20.7% to 61.2%,
respectively, in reagent and source-quality water samples
(10L) spiked with approximately 100 oocysts [2]. The
other laboratory obtained an average 49% recovery
when 100L volumes of tap water were spiked with 51
(four replicates) or 16 (six replicates) oocysts [3]. When
molecular biology-based methods such as RT-PCR and
cell culture were used as part of this protocol, one viable
oocyst per 60L of environmental water was detected [4].
However, the genus-specific character of these C. parvum
hsp70 primers has been questioned [5]. Using the US
EPA method 1623, in conjunction with 18S rRNA gene
nested PCR, Sturbaum et al. [6] reported successful
ARTICLE IN PRESS
*Corresponding author. Tel.: +1-301-504-6443; fax: +1-
301-504-6608.
E-mail address: jhiggins@anri.barc.usda.gov (J.A. Higgins).
0043-1354/03/$-see front matter. Published by Elsevier Science Ltd.
doi:10.1016/S0043-1354(03)00251-3

Page 2

detection of as few as five oocysts spiked into 1L of
source water.
Other recovery and detection methods that have been
evaluated include membrane filtration and flocculation
[1], membrane filter dissolution [7–9], and continuous
flow centrifugation (CFC) [10]. When 45L volumes of
source water were spiked with 2.6?104oocysts, the
mean recovery for three replicates was 26.8%, compared
to 0.5% for cartridge filtration and 18.7% for calcium
carbonate flocculation [10]. Swales and Wright [11] also
compared CFC with cartridge filtration, on 100L water
samples with nephelometric turbidity unit (NTU) values
of 1–5. They spiked the water with 10 oocysts/L. Mean
recoveries for the CFC method were 14.8% and 13.0%
for the 1 and 5NTU samples, respectively, compared to
9.7% and 9.0% for cartridge filtration, respectively.
More recently, Borchardt and Spencer [12] used blood
cell separator centrifuges and reported recovery efficien-
cies of 78% when 500 oocysts were spiked into 25L of
pond water, and 97% when 600 Giardia cysts were
spiked into 30L of pond water.
These results suggested that CFC offered recoveries
equivalent to cartridge filtration, with improved econo-
my and ease of use compared to filter-based methods.
We report here on our evaluation of CFC for recovering
Cryptosporidium oocysts, and Giardia intestinalis cysts,
from spiked source water samples.
2. Materials and methods
2.1. Experimental design and collection of water samples
The term ‘‘run’’ is used to refer to the entire assay in
which the spiked water sample is processed using CFC,
subjected to overnight bead capture, and subsequent
IFA microscopy the following morning.
The term ‘‘oocyst recovery’’ refers to the number of
oocysts detected following a CFC run. The term
‘‘method sensitivity’’ refers to the percentage of CFC
runs that were successful in detecting spiked oocysts.
The experimental design was as follows: in the first
series of experiments, 10 source water samples (10L)
were spiked with 100 oocysts of Cryptosporidium parvum
oocysts, subjected to CFC and IMS, and the pellet of
recovered oocysts was enumerated using immunofluor-
escence antibody (IFA) staining and microscopy. In the
second series of experiments, 10 source water samples
(10L) were spiked with 100 oocysts, subjected to CFC,
and the recovered oocysts assayed by nested PCR for
two C. parvum genes. These two series of experiments
were performed in 2001. A third series of experiments,
taking place in the Spring and Summer of 2002, was
conducted on source water samples (10L) spiked with 10
oocysts. Following CFC and IMS, the recoveries of
oocysts and cysts were determined by immunofluores-
cence IFA microscopy. For seven of these 12 runs
Giardia cysts were also spiked into the 10L of source
water; for these ‘‘dual spiked’’ samples, the first over-
night IMS incubation was done using the Aureon anti-
Giardia kit. The 10mL, post-bead capture sample was
then subjected to overnight incubation using the Dynal
anti-Cryptosporidium kit, with 1mL each of buffers A
and B simply being added to the sample, along with the
required 100mL of anti-Cryptosporidium Dynalbeads.
A fourth series of experiments used 10 fluorescence-
activated cell sorter (FACS)-enumerated oocysts as a
spiking dose in 10L of tap water. During Fall 2002, a
total of 10 runs were done, with one unspiked control
run done at the midway point. The oocyst pellets from
this series of experiments were subjected to overnight
IMS and IFA microscopy.
Source water samples were obtained from facilities at
theWashington Suburban
(WSSC) in Laurel, Maryland, and from Beaver Dam
Creek (BDC) located on the campus of the Beltsville
Agricultural Research Center at Beltsville, MD. Car-
boys (20L) of water were stored at 4?C until used (later
changed to room temperature; once the CFC protocol
had been finalized it was routine to assay 60–80L of
water within a 1-week period). For experiments using
tap water, 10L volumes were obtained from taps located
in building 173 of the Beltsville Agriculture Research
Service campus.
SanitaryCommission
2.2. Oocyst enumeration
Cryptosporidium parvum oocysts of the Beltsville-1
strain, as well as Giardia intestinalis cysts, were
recovered from the feces of experimentally infected
calves [13]. Briefly, the fecal material was sieved to
remove larger particulates and then subjected to cesium
chloride gradient centrifugation in a 250mL centrifuge
tube, at 250?g for 1h at 5?C. The layer containing the
oocysts (approximately 10–20mL) was transferred to a
250mL centrifuge tube, deionized water was added to
bring the volume to 250mL, and the tube was
centrifuged at 250?g for 15min at 15?C. The resultant
pellet was washed twice with deionized water, and
resuspended in a volume of 100–200mL. Oocysts used
for spiking were stored for no longer than 8 weeks after
isolation.
Enumeration of oocysts used to create stock concen-
trations was accomplished by pipetting 10mL of the
oocyst suspension onto a Neubauer hemacytometer and
counting them under bright field microscopy. A single
count consisted of the mean of four different counts at
the outer 4?4 grid of the hemacytometer. Counts from
the hemacytometer method were used to prepare stock
concentrations of oocysts at 1?106/mL.
For experiments in which oocysts were spiked into
10L volumes of tap water, 10-oocyst aliquots in 10mL of
ARTICLE IN PRESS
J.A. Higgins et al. / Water Research 37 (2003) 3551–3560
3552

Page 3

sterile water (enumerated using a fluorescence activated
cell sorter, or FACS) were purchased from the
Wisconsin State Laboratory of Hygiene Environmental
Health Division Flow Cytometry laboratory (Madison,
WI).
To count oocysts recovered from continuous flow
centrifugation experiments, as well as dilutions of
oocysts used for source water spiking experiments,
1mL of oocyst suspension was mixed with 1mL of
Merifluort immunofluorescence reagent (Meridian Di-
agnostics, Cincinnati, OH), pipetted onto each well of a
3-well Teflon-coated microscope slide (Cell-linet, Erie
Scientific, Portsmouth, NH), and examined at 400?
magnification (this emendation of the Merifluor proto-
col was done because in our experience, performing the
stain washing step while oocysts are on the microscope
slide can sometimes result in loss of oocysts and
consequently lead to erroneous counts). The slides were
examined via IFA microscopy by the same two
individuals throughout the experiment to ensure con-
sistency in oocyst identification and enumeration.
2.3. Continuous flow centrifugation/immunomagnetic
separation
A photograph of the setup for the CFC is shown in
Fig. 1A and a flow chart of the procedure is given in
Fig. 1B. Oocysts were first spiked into 50mL of sample
water in a conical centrifuge tube, vortexed, and then
deposited into the carboy containing the remaining
9.95L of water. The carboy was shaken and then the
contents pumped, using a MasterFLex peristaltic pump
(Cole-Parmer, IL), through a No. 11 Model continuous
flow centrifuge (Lavin, Hataboro, PA) at a rate of
200mL/min, at a speed of 7140rpm (4200?g). The
Lavin CFC is 11.5in (29.2cm) high, 12in (30.4cm) in
diameter, and weighs 35lbs (15.8kg). The body and lid
are aluminum, and the rotor is stainless steel. The
maximum rpm is 10,000 (8550?g). A diagram of the
Lavin CFC operating principle and other pertinent
information is available at: www.lavincentrifuge.com.
After passage of the entire 10L of oocyst spiked source
water, 1L of tap water was added to the carboy, shaken
to clean the sides and top of the carboy, and pumped
through the centrifuge. Pelleted material (approximately
130mL) was collected from the bowl of the CFC and
transferred to a 200mL polypropylene conical centrifuge
tube. The bowl was then washed with approximately
30mL of water, which was added to the 200mL conical
tube; the combined eluate and wash volume of 160mL
was brought up to 200mL with sterile water, and the
tube was centrifuged in a swinging bucket rotor at
2800rpm (1500?g) for 20min. All but approximately
5mL of supernatant was aspirated and discarded. The
pellet was then resuspended in the 5mL of supernatant
and transferred to a 15mL polypropylene conical tube;
10mL of sterile water were used to rinse the 200mL
tube, and this rinse then added to the 15mL tube, which
was then centrifuged in the swinging bucket rotor at
1500?g for 20min. Fourteen milliliter of supernatant
was aspirated and discarded. For those samples under-
going IMS, the final recovery volume of approximately
1mL was brought up to 10mL with sterile water and
subjected to IMS, using the Dynal Dynabeadssanti-
Cryptosporidium kit (Lake Success, NY) and the Aureon
anti-Giardia IMS kit (ImmTech, Inc. New Windsor,
MD), according to the manufacturer’s instructions, save
that bead incubation was performed overnight rather
than for 1h.
Unspiked control runs—consisting of either 10L tap
water or 10L of source water—were performed to
monitor for false-positive results due to the recovery of
oocysts and cysts retained in the CFC apparatus (refer
to Tables 2 and 3).
In between runs the CFC rotor was washed using a
scrubbing brush and a soap solution and rinsed several
times with tap water. The carboys used to deliver the
spiked 10L of source water were also cleaned with a
soap solution and rinsed several times with tap water.
2.4. Instagenet DNA extraction and PCR
For the Dynal bead–oocyst pellets, a DNA extraction
procedure using Instagenet matrix (Bio-Rad, Hercules,
CA) was performed. Briefly, the pellet (15–20mL) in a
1.5mL microfuge tube was subjected to two freeze–thaw
cycles in a dry ice and methanol bath and 55?C water
bath, then 100mL of Instagene matrix was added and the
tube incubated at 56?C for 15min, followed by an
incubation at 100?C for 8min. The tube was then
centrifuged to pellet the matrix and 10mL of supernatant
used as template for PCR.
For the Cp41 gene [14] Genbank Accession No.
AF144621; the outer forward primer was Cp41OF: 50
GAG GAG ATG GAC TAT TCT AGG; outer reverse
primer Cp41OR: 50GCA ACA GTA GTA AGA GTG
GTA; inner forward primer Cp41 IF: 50TGT ATG
AAT TGG ATA TAT TAT TA; and inner reverse
primer Cp41 IR: 50GTA AAA GCA ACA CCA TTA
CTA. The TRAP C-1 primers Cp.Z, Cp.W, and Cp.E
were used according to the protocol of Spano et al. [15]
Cp.E and Cp.Z served as outer primers, and Cp.E and
Cp.W as inner primers.
PCR was done in 50mL volumes, containing 1U Taq
polymerase (Life Technologies, Gaithersburg, MD),
200mM each dNTP, 1.5mM MgCl2, 5mL 10? PCR
buffer, and 50pmol of each primer. Cycling parameters
for the Cp41 gene PCR were: 95?C for 1min, followed
by 35 cycles of 95?C for 15s, 50?C for 30s, and 60?C for
1min. Cycling parameters for the TRAP C-1 nested
PCR followed the published protocol (Spano et al.,
ARTICLE IN PRESS
J.A. Higgins et al. / Water Research 37 (2003) 3551–3560
3553

Page 4

ARTICLE IN PRESS
Fig. 1. Panel A: continuous flow centrifugation (CFC) apparatus. A 10L volume of source water in a plastic carboy (A) is pumped via
a peristaltic pump (B) into a Lavin CFC (C) at a rate of 200mL/min. Overflow from the rotor is discarded in another carboy (D). Panel
B: flowchart depicting the steps in the CFC procedure.
J.A. Higgins et al. / Water Research 37 (2003) 3551–3560
3554

Page 5

1998), albeit with the PCR extension step time increased
to 2min.
3. Results
3.1. Oocyst enumeration
IFA microscopy examination of oocyst preparations
used in spiking experiments showed that the observed
counts were not markedly different than expected counts
(i.e., the coefficient of variation did not exceed 10%). An
example of one series of counts is provided in Table 1.
Here, each replicate consists of 10, 1mL aliquots with an
expected content of 10oocysts/aliquot. The average
count of oocysts per sum of 10, 1mL aliquots was
97.378.0, with a range of 85–111 (CV=8.0%). ANO-
VA performed on the data in this table indicated no
significant difference in oocyst counts between replicates
(F ¼ 0:6271oF0:05 9;901:99;P ¼ 0:771). We interpreted
these results to indicate that our spiking doses did not
contain significantly more oocysts than predicted from
enumeration data.
3.2. Optimization of CFC parameters
A series of experiments was conducted using a spiking
dose of 10,000 oocysts in 10L of tap water. Percent
recovery for 200mL/min at 2800?g was 7%, for
200mL/min at 4200?g, 19%; for 250mL/min at
4200?g, 5%; for 200mL/min at 4200?g, 27.3%; and
for 200mL/min at 6000?g, 20%. Consequently, CFC
runs used a flow rate of 200mL/min at 4200?g (note
that on the Lavin CFC this is equivalent to a dial setting
of ‘‘6’’).
3.3. Recovery of oocysts from spiked water samples using
CFC
In general, it took 2.5h to perform the CFC
procedure, and another 12 (i.e., overnight) to conduct
the IMS separation. Subsequent IFA microscopy on
purified oocysts usually took under 1h. DNA extraction
using Instagene matrix, followed by nested PCR, added
another 4h to the 14.5h devoted to CFC and IMS.
CFC pellets derived from source water were essen-
tially sediment and ranged in size from 0.5 to 1.0mL.
NTU values for WSSC source water samples ranged
from 1.1 to 3.9, except for one source water sample that
had a value of 22. Readings for BDC source water were
higher and ranged from 6.1 to 27.0NTU (however, a
collection taken on February 13, 2002 was particularly
turbid due to inadvertent stirring up of sediment and
readings of 134, 135, and 133NTU were recorded). We
observed very high NTU readings (>500) on the pellet
of sediment recovered by the CFC, even when mixed
with 10mL of water and 1mL each Dynal kit buffers A
and B for the initial IMS step; for example, of six
samples of WSSC-derived source water assayed in June,
2001, the average NTU reading for the resuspended
pellets was 982 (737.7, range 916–1014).
Preliminary experiments, consisting of three assays in
which 1?103oocysts were spiked into 10L of WSSC
source water, showed recoveries of 174, 115, and 135
oocysts (17%, 11%, and 13.5%, respectively), as
determined by IFA microscopy. The success of the
CFC technique in detecting these spiking doses led us to
attempt recovery of smaller numbers of oocysts.
Accordingly, 10 runs were performed in which 100
oocysts were spiked into 10L volumes of source water
from the WSSC; oocysts were recovered using CFC and
an overnight Dynal bead IMS step; and IFA microscopy
was used to detect recovered oocysts. Another 10 runs
ARTICLE IN PRESS
Table 1
Enumeration of Cryptosporidium parvum oocysts used in source water spiking experiments
Rep# Well# Total
12345678910
1
2
3
4
5
6
7
8
9
11
5
11
10
13
8
13
9
8
11
16
9
10
10
11
14
10
14
8
10
9
6
9
7
11
12
9
5
8
11
10
2
6
6
2
8
9
5
4
9
9
8
1089
9
16
13
9
18
8
12
8
3
12
10
97
87
98
104
97
111
99
85
104
91
x ¼ 97:378:01
8
4
8
13
13
11
14
8
7
11
11
7
15
15
7
10
9
9
12
10
11
10
10
15
11
9
14
810
15
9
9
13
3
14
10
8
9
13
9
8
9
16
107
Ten, 1mL aliquots of oocysts were examined using MeriFluor IFA staining with 400? magnification.
J.A. Higgins et al. / Water Research 37 (2003) 3551–3560
3555