Microbiology Papers in Press. Published February 8, 2015 as doi:10.1099/mic.0.000051
Microbiology Ecology and Genomic Features of Infection with Mycobacterium avium subspecies paratuberculosis in Egypt. --Manuscript Draft-Manuscript Number:
MIC-D-14-00005R2
Full Title:
Ecology and Genomic Features of Infection with Mycobacterium avium subspecies paratuberculosis in Egypt.
Short Title:
Ecology and Genomics of M. paratuberculosis
Article Type:
Standard
Section/Category:
Environmental Biology
Corresponding Author:
Adel M. Talaat, Ph.D. University of Wisconsin-Madison Madison, WI UNITED STATES
First Author:
Adel S. Amin
Order of Authors:
Adel S. Amin Chung-Yi Hsu Samah F. Darwish Pallab Ghosh Eman M. Abdul-Fatah Tahani S. Behour Adel M. Talaat, Ph.D.
Abstract:
Mycobacterium avium subspecies paratuberculosis (M. paratuberculosis) is the causative agent of paratuberculosis, or Johne's disease, in cattle with potential involvement in cases of Crohn's disease in humans. Johne's disease is found worldwide and is economically important for both beef and dairy industries. In an effort to characterize this important infection in Egypt, we analyzed the ecological and genomic features of recent isolates of M. paratuberculosis. In this report, we examined 26 Holstein dairy herds distributed throughout Egypt from 2010 to 2013. Using PCR analysis of fecal samples, we estimated an average herd level prevalence of 65.4% with animal level infection that reached an average of 13.6% among animals suffering from diarrhea. Whole genome sequencing of field isolates identified numerous single nucleotide polymorphisms among field isolates relative to the standard M. paratuberculosis K-10 genome. Interestingly, the virulence of M. paratuberculosis isolates from Egypt revealed diverse virulence phenotypes in the murine model of paratuberculosis with significant differences in tissue colonization, particularly during the chronic stage of infection. Overall, our analysis confirmed that Johne's disease is a newly identified problem in Egypt and indicated that M. paratuberculosis has potentially diverse genotypes that impact its virulence. Further ecological mapping and genomic analysis of M. paratuberculosis will continue to enhance our understanding of the transmission and evolutionary dynamics of this pathogen under natural field conditions.
Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation
Manuscript Including References (Word document) Click here to download Manuscript Including References (Word document): Talaat etal ver2.5.docx
1
Ecology and Genomic Features of Infection with Mycobacterium avium subspecies
2
paratuberculosis in Egypt. Adel S. Amin1, Chung-Yi Hsu2, Samah F. Darwish1, Pallab Ghosh2, Eman M.
3
Abdul-Fatah1, Tahani S. Behour1, Adel M Talaat2,3.
4 5
1
Biotechnology Research Unit, Animal Reproduction Research Institute (ARRI), Giza, Egypt;
6
2
Department of Pathobiological Sciences, University of Wisconsin-Madison, Madison, WI,
7
USA; 3 Department of Food Hygiene and Control, Faculty of Veterinary Medicine Cairo
8
University, Giza, Egypt.
9 10 11
#
12
Adel M. Talaat, M.V.Sc., Ph.D.
13
Laboratory of Bacterial Genomics, Department of Pathobiological Sciences, University of
14
Wisconsin-Madison, 1656 Linden Drive, Madison, WI 53706
15
Email:
[email protected] 16
Tel. (+1) 608 262 2861
17
Fax: (+1) 608 262 7420
Address correspondence to:
18 19
Running title: Ecology and Genomics of M. paratuberculosis.
20 21
Genome Accession Numbers: CP010113, CP010114
22
1
23 24 25
Summary:
26
agent of paratuberculosis, or Johne’s disease, in cattle with potential involvement in cases of
27
Crohn’s disease in humans. Johne’s disease is found worldwide and is economically important
28
for both beef and dairy industries. In an effort to characterize this important infection in Egypt,
29
we analyzed the ecological and genomic features of recent isolates of M. paratuberculosis. In
30
this report, we examined 26 Holstein dairy herds distributed throughout Egypt from 2010 to
31
2013. Using PCR analysis of fecal samples, we estimated an average herd level prevalence of
32
65.4% with animal level infection that reached an average of 13.6% among animals suffering
33
from diarrhea. Whole genome sequencing of field isolates identified numerous single nucleotide
34
polymorphisms among field isolates relative to the standard M. paratuberculosis K-10 genome.
35
Interestingly, the virulence of M. paratuberculosis isolates from Egypt revealed diverse
36
virulence phenotypes in the murine model of paratuberculosis with significant differences in
37
tissue colonization, particularly during the chronic stage of infection. Overall, our analysis
38
confirmed that Johne’s disease is a newly identified problem in Egypt and indicated that M.
39
paratuberculosis has potentially diverse genotypes that impact its virulence. Further ecological
40
mapping and genomic analysis of M. paratuberculosis will continue to enhance our
41
understanding of the transmission and evolutionary dynamics of this pathogen under natural field
42
conditions.
Mycobacterium avium subspecies paratuberculosis (M. paratuberculosis) is the causative
43
2
44 45 46
Introduction.
47
causative agent of paratuberculosis or Johne's disease (Collins, 1996). Johne’s disease is a
48
chronic debilitating infection that affects a wide range of animals, including both wild and
49
domestic ruminants (Lombard, 2011a). Two major types of M. paratuberculosis infect
50
ruminants: the S-type, Type I mainly infecting sheep; and the C-type, Type II mainly infecting
51
cattle and other ruminants. Recently, a sub-lineage of the S-type (Type III) has emerged in sheep
52
and camels (Ghosh et al., 2012). Only molecular approaches can differentiate between the strain
53
types. Clinically infected cows can shed 106–108 CFU/g in fecal materials and as little as 103
54
CFU/animal is sufficient to infect other animals through the fecal-oral route (Whittington &
55
Sergeant, 2001). Shedding of M. paratuberculosis in feces of infected animals is intermittent
56
(Raizman et al., 2009) complicating studies to estimate prevalence and detect the presence of
57
infection. At the late stage of infection, animals suffer from chronic diarrhea, weight loss, low
58
milk yield and low mortality which results in significant economic losses to the dairy industry in
59
the USA (Lombard, 2011a) and worldwide (Barrett et al., 2006). Globally, the prevalence of the
60
disease can range from 3-4% of the herds in regions with low incidence, such as England
61
(Cetinkaya et al., 1996), to more than 90% of dairy operations in regions of high incidence, such
62
as the USA (Lombard et al., 2013). Although culturing of fecal samples is considered the gold
63
standard for the diagnosis of Johne’s disease in living animals (Shin et al., 2007), more recent
64
studies relied on fecal PCR amplification (Sonawane & Tripathi, 2013) to overcome problems
65
associated with culturing (e.g. low sensitivity, long periods of incubation). The major objectives
66
of this report are to examine the herd level prevalence of Johne’s disease in an environment
Mycobacterium avium subspecies paratuberculosis (M. paratuberculosis), is the
3
67
where no control strategy has been implemented and to characterize the virulence of M.
68
paratuberculosis isolates with known genetic polymorphisms.
69
In Egypt, statistics about the prevalence and economic impact of Johne’s disease are
70
scarce. In one report, Johne’s disease was confirmed in 47% of dairy cows with diarrhea from 5
71
governorates in Egypt (Salem et al., 2005). Surprisingly, even healthy animals showed a high
72
level of prevalence (27%) (Salem et al., 2005), indicating the potential underestimation of this
73
disease in Egypt. Despite this earlier report and observations from several field veterinarians, no
74
control measures have been taken to control JD in this country. With a sensitivity level no more
75
than 50% for serum ELISA (Collins et al., 2005) or 76% for fecal PCR in high shedder animals
76
(Wells et al., 2006), most of the studies underestimate the true prevalence of JD. Nevertheless,
77
some of the PCR-based techniques, when combined with other molecular approaches, could be
78
used to genotype strains of M. paratuberculosis (e.g. short sequence repeats) (Castellanos et al.,
79
2012; Pradhan et al., 2011) to better track the source of infection. Based on IS900 (Stevenson et
80
al., 2002), IS1311 (Whittington et al., 1998) and gyrB sequence polymorphisms (Castellanos et
81
al., 2012), the major groups of M. paratuberculosis strains were identified along with host
82
associations (Type I for sheep and Type II for cattle). Fortunately, genome-wide approaches
83
could improve the sensitivity and specificity of assays used to detect Johne’s disease
84
(Castellanos et al., 2012). Recently, several clinical isolates of M. paratuberculosis were
85
subjected to whole genome sequencing which provided an accurate estimate for strain
86
geographical and host association (Ghosh et al., 2012; Hsu et al., 2011; Wu et al., 2006) with
87
implications on our understanding of strain evolution (Bannantine et al., 2012). Genome-based
88
techniques could also provide more targets for developing better markers for tracing the source
89
for M. paratuberculosis infection.
4
90
In this study we employed fecal PCR on suspected cases with diarrhea collected from 13
91
governorates throughout Egypt to examine the ecology of M. paratuberculosis infection. Our
92
analysis indicated a herd level prevalence of Johne’s disease that reached 65.4%. Culturing of
93
fecal PCR-positive samples resulted in the isolation of 8 strains of M. paratuberculosis that were
94
genotyped to belong to M. paratuberculosis–C Type (Type II). Full genome sequencing of two
95
isolates, MAP-E1 and MAP-E93, predicted over 8 times as many single nucleotide polymorphic
96
(SNPs) sites in the genome of strainE1 than in E93 compared to the standard M.
97
paratuberculosis K10 genome. Interestingly, significant differences were identified in the
98
virulence phenotype associated with each isolate in the murine model of paratuberculosis, most
99
likely related to these genotypic differences.
100
Materials and Methods:
101
Field study and sampling strategy. A total of 1290 among 2576 Holstein cows 24 months or
102
older with chronic diarrhea suspected of contracting Johne’s disease were enrolled in this study.
103
Enrolled animals were raised on one of 26 dairy farms in 13 different governorates representing
104
different geographical regions within Egypt (Fig. 1). Fecal samples from all 1290 animals were
105
screened using PCR to amplify IS900 target as detailed before (Hines et al., 2007). Serum
106
samples were collected from a subgroup of these animals (N=120) for further follow up during
107
the 3 year period of the study. All serum samples were tested for the detection of antibodies
108
against M. paratuberculosis using the ParaChek ELISA kit as per manufacturer’s directions
109
(Biocor Animal Health Inc. Omaha, NE) (Wells et al., 2006). Additionally, a total of 334 and 80
110
fecal samples were collected from 2 Holstein dairy herds in Giza and Behaira governorates,
111
respectively, where all cows ≥24 months old were sampled for fecal PCR analysis.
112
Culturing fecal samples. Farm veterinarians collected fecal samples directly from the rectum of
113
the examined animal. Before culturing, Hexadecyl pyridinium chloride (HPC) (Sigma/Aldrich, 5
114
St. Louis, Missouri, USA) was used to decontaminate each PCR positive fecal samples using a
115
protocol developed before (Wells et al., 2002), with some modifications. Briefly, 3 grams of
116
each fecal sample were added to 30 ml of 1% HPC and incubated for 18 h at ambient
117
temperature. The upper liquid layers were centrifuged (3,200 g for 20 min) and pellets
118
resuspended in 0.5 ml of a buffer composed of 0.1 M PBS (phosphate buffered saline, 136.9 mM
119
NaCl, 1.46 mM KH2PO4, 8.1 mM Na2HPO4 x 2H2O, and 2.7 mM KCl, pH 7.4 in the presence of
120
100 μg/ml vancomycin, 100 μg/ml naladixic acid and 50 μg/ml amphotericin B (Sigma-Aldrich,
121
St. Louis, MO). A 0.2 ml aliquot of each final fecal suspension was used for inoculating culture
122
media. All prepared samples were cultured on 2 tubes of modified Herrold’s Egg Yolk Media
123
(HEYM) with and without mycobactin J (Allied Monitor, Inc., Fayette, Missouri, USA) as well
124
as Middlebrook 7H10 agar plates supplemented with mycobactin J as outlined before (Hines et
125
al., 2007). Inoculated tubes were incubated at 37°C for 5 months and examined weekly.
126
Additionally, the Ziehl–Neelsen stain (ZNS) was used to microscopically evaluate colonies
127
resembling M. paratuberculosis . Positive ZNS samples were tested by PCR for confirmation.
128
Polymerase chain reaction assay. All fecal samples were prepared for PCR amplification as
129
described before (Stabel et al., 2004) with some modifications. Briefly, each fecal sample (3 g)
130
was diluted in 9 ml of Tris-ethylenediamine tetraacetic acid (TE) buffer (10 mM Tris–HCl, 1
131
mM EDTA; pH 7.6), centrifuged at 200 g for 30 sec, and the supernatant was diluted 1:100 in
132
TE buffer. One milliliter of each dilution was centrifuged at 13,000 g for 2 min to obtain pellets
133
that were resuspended in 500 µl of TE buffer and boiled for 10 min. After cooling to room
134
temperature, 4 µl of RNase (500 mg/ml) were added to each sample and a total of 5 µl from each
135
preparation was used for the PCR. For colony PCR, a quick protocol for DNA extraction was
136
used as detailed before (Talaat et al., 1997). The target for PCR amplification included a 229 bp
6
137
fragment in the IS900 sequence with specific forward (5`-CCG CTA ATT GAG AGA TGC
138
GAT TGG-3`) and reverse primers (5`-AAT CAA CTC CAG CAG CGC GGC CTC G-3`) as
139
described previously (Hsu et al., 2011). Amplification conditions included an initial
140
denaturation at 94°C for 3 min followed by 50 cycles of denaturation at 94°C for 60s, annealing
141
at 60°C for 60s and extension at 72°C for 60s. A final extension step at 72°C for 7 min was also
142
allowed.
143
Genotyping of M. paratuberculosis isolates. All M. paratuberculosis isolates were subjected to
144
genotyping using a protocol based on short sequence repeats (SSR) (Kasnitz et al.).
145
Representatives of cattle (K-10) and sheep (S329) M. paratuberculosis strains were included in
146
the genotype analysis. The loci for SSR 1-6, and 8-11 were used for genotyping using primers
147
published before (Amonsin et al., 2004). PCR reactions consisted of 25 µl, containing 1 M
148
betaine, 50 mM potassium glutamate, 10 mM Tris-HCl pH 8.8, 0.1% Triton X-100, 2 mM
149
magnesium chloride, 0.2 mM dNTPs, 0.5 mM of each primer, 0.5 U of Taq DNA polymerase
150
(Promega, Madison, WI) and 5 µl of genomic DNA. The amplification conditions included an
151
initial denaturation step of 94°C for 5 min followed by 35 cycles of denaturation at 94°C for 30
152
s, annealing at 55°C for 30 s and extension at 72°C for 1 min, and followed by a final extension
153
at 72°C for 7 min (Hsu et al., 2011). Following PCR, amplicons were cleaned up and processed
154
for sequencing with the BigDye® Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems,
155
Foster City, CA) (Hsu et al., 2011). Each amplicon was sequenced from both directions to
156
confirm repeats of each SSR site.
157
For differentiating Type I and Type II isolates, the IS1311 sequence was targeted with gene-
158
specific primers (forward 5` GCG TGA GGC TCT GTG GTG AA 3` and reverse 5` TCA GAG
159
ATC ACC AGC TGC AC 3`). PCR products were evaluated by electrophoresis in 2% agarose
7
160
gels pre-stained with ethidium bromide (0.5 mg/ml). The resulting single-band products of
161
interest were purified and extracted using the Wizard® SV Gel and PCR Clean-Up System
162
(Promega). A total of 250–300 ng of DNA and 1 U of the restriction enzymes HinfI (NEB,
163
Beverly, MA) were used in the digestion reactions according to the manufacturer’s
164
recommendations. Digestion reactions were assessed using 3% agarose gels stained with
165
ethidium bromide.
166
Next-generation whole genome sequencing. Purified genomic DNA (1-5 μg) isolated from
167
each target strain were sent to the UW-Madison Biotechnology Center DNA Sequencing Facility
168
for whole genome sequencing using the Illumina HiSeq 2000 platform. Assembly and analysis
169
of sequencing data were performed with the CLC Genomics Workbench 4.6 (CLCBio,
170
Cambridge, MA). Raw data from two Egyptian cattle isolates, E1 and E93, was assembled
171
against the revised reference strain M. paratuberculosis K10 (Wynne et al., 2010) using CLCBio
172
Genomic Workbench software (version 5.1). The overall percent homology between genome
173
sequences was calculated based on the consensus length of each sequenced genome divided by
174
the calculated length of the reference strain and then multiplied by 100. Additionally, the
175
MAUVE algorithm was used to align paired or multiple genomes for comparative purposes, as
176
outlined before (Darling et al., 2004; Perna et al., 1998). The gapped consensus sequence of
177
each strain was imported to MAUVE for sequence alignment at the default seed weight setting.
178
SNP detection was also performed using the software with a presence frequency equal to or more
179
than 50%. SNPs were used for the phylogenetic analysis. For each genome, loci containing
180
SNPs in E1, E93, and our other previously sequenced strains (Hsu et al., 2011) were tabulated
181
using in-house software. The concatenated SNP files for each strain were aligned with MEGA 6
8
182
(Tamura et al., 2007) and used to build a neighbor-joining phylogenetic tree with 1000 bootstrap
183
replications.
184
Murine infection with M. paratuberculosis. BALB/c mice at 3 to 4 weeks of age were housed
185
in a pathogen-free environment for at least 2 weeks before infection. Groups of BALB/c mice (n
186
= 15) were infected with M. paratuberculosis Egyptian strains (E1 and E93) by intraperitoneal
187
injection as outlined before (Shin et al., 2006; Wu et al., 2007). As controls, groups of mice
188
were inoculated with M. paratuberculosis JTC1285 and K-10 isolates previously analyzed in our
189
laboratory (Hsu et al., 2011). All M. paratuberculosis strains used in this study have been
190
passaged fewer than 5 times from the original stock. For animal infection, M. paratuberculosis
191
strains were cultured to mid-log phase (optical density at 600 nm of 1.0) and centrifuged at 5,000
192
rpm for 10 min to harvest mycobacterial cells before resuspension in PBS. Each mouse was
193
injected intraperitoneal with approximate 108 M. paratuberculosis CFU (approximately 0.5
194
ml/mouse). Infected mice (n = 5) were sacrificed at 3, 6, and 12 weeks post infection, and livers,
195
spleens, intestines, and mesenteric lymph nodes (MLN) were collected for both bacterial counts
196
and histopathological examination. Tissue sections collected for histopathology were preserved
197
in 10% neutral buffered formalin before being embedded in paraffin, cut into 4 to 5 µm sections,
198
and stained with hematoxylin and eosin or the Ziehl-Neelsen method for acid-fast staining
199
(Ghosh et al., 2013). The severity of the inflammatory response in the tissue sections was ranked
200
using a score of 0 (lowest) to 5 (highest) based on lesion size and number per field. All statistical
201
tests of histological scores or colony forming units to compare between groups were conducted
202
in Microsoft Excel using Student’s t-Test where a p98% homology to the genome of the reference
256
strain. Zero coverage regions were observed in 46 and 216 loci in E1 and E93 genomes,
257
respectively. Summary statistics of the reference assembly for both genomes are compiled in
258
Table 3. To examine large-scale genomic rearrangements, we used the default alignment
259
algorithm found in the software package MAUVE which was developed to compare the genomes
260
of closely related isolates (Darling et al., 2004). In this analysis, no regions of insertions or
261
deletions were identified, only single nucleotide polymorphism (SNPs) (Fig. 3).
262
Interestingly, the number of predicted SNPs for the E1 consensus sequence, 432 SNPs, was
263
much higher than that predicted for the E93 sequence, 50 SNPs, when the genomes of both
264
isolates were aligned in reference to the standard M. paratuberculosis K-10 strain (Table 3). The
265
majority of SNPs (>50%) are non-synonymous (nSNP), suggesting a positive selective pressure
266
on the identified genes. Only 12 SNPs are common in both E1 and E93 with the same variant
267
nucleotide; among these, 9 are nSNPs. Changes at the loci of all 12 shared SNPs have been seen
268
in other sequenced Type II M. paratuberculosis isolates from locations around the globe (Hsu et
269
al., 2011) and 4 of these SNPs appeared to have the same nucleotide inversion among all the
270
sequenced Type II M. paratuberculosis strains (Table 4). Interestingly, we observed a few SNPs
271
in the PPE and PKS protein families, expected to modulate antigen presentation. Additionally,
12
272
more SNPs were seen in members of the mce (mammalian cell entry) gene family in both E1 and
273
E93 genomes compared to other genes. Based on phylogenetic analysis of SNPs from both
274
strains and others sequenced before by our group (Hsu et al., 2011), we confirmed the
275
divergence of E1 and E93 strains (Fig. 4). The whole genome sequences for both isolates
276
(CP010113, CP010114) are deposited in GenBank.
277
M. paratuberculosis virulence in murine models.
278
To evaluate the virulence of the M. paratuberculosis E1 and E93 strains recently isolated
279
from Holstein dairy herds, we used the murine model of paratuberculosis (Shin et al., 2006).
280
Although mice do not suffer from diarrhea following M. paratuberculosis infection, colonization
281
of liver, spleen, and intestine is usually a good indicator for disseminated infection (Hines et al.,
282
2007; Shin et al., 2006). In addition, the presence of a granulomatous reaction (poorly defined
283
aggregates of lymphocytes and macrophages harboring mycobacteria) is the most characteristic
284
lesion associated with mycobacterial infection in this model (Hines et al., 2007). In this
285
experiment, each mouse group received approximately 108 CFU/animal using intraperitoneal
286
injection and organs were harvested at 3, 6, and 12 weeks post infection. These time points were
287
chosen based on a previous study that found signs of progressive infection in mice at these time
288
points following intraperitoneal inoculation (Shin et al., 2006). The E1 strain, with the highest
289
number of SNPs, displayed a similar colonization pattern to the laboratory grown strain,K10, in
290
liver and intestine at all time points (Fig. 5). On the other hand, E93 showed a similar
291
colonization pattern to the Type II clinical isolate, JTC1285 (Hsu et al., 2011). The disparity in
292
colonization patterns was also noted in spleen and mesenteric lymph nodes of infected animals
293
(Fig. 5 C, D). The later 2 organs were not collected from mice infected with K-10 and JTC1285
294
isolates.
13
295
Histologically, all infected animals showed early granulomatous inflammation
296
(lymphocyte aggregates with a few macrophages) in liver and spleen starting from 3 weeks post
297
challenge (Fig. 6), consistent with the characteristics of the murine model of paratuberculosis.
298
Acid fast-bacilli (AFB) were noticeable only in livers of mice inoculated with E93 but not in
299
livers of mice inoculated with E1. As expected, by 12 weeks post infection, the granulomatous
300
inflammation increased in size and number as the infection progressed. In general, the
301
inflammatory response and AFB were more noticeable in livers and spleens of mice inoculated
302
with the E93 compared to E1. Interestingly, Peyer’s patches (organized lymphoid organs found
303
in the intestine) were engorged with lymphocytes and more noticeable in mice inoculated with
304
the E1 isolate compared to those inoculated with E93 isolate starting from 6 weeks onward.
305
Despite using the intraperitoneal model of murine paratuberculosis, the observed signs of
306
disseminated infection (colonization and lesions in liver and spleen) provided evidence for
307
systemic infection and proper inoculation of M. paratuberculosis strains used in the study. Such
308
observations were consistent with earlier characterization of this model in our hands (Hines et
309
al., 2007).
310
Discussion.
311
Earlier reports of Johne’s disease in Egypt indicated the presence of this disease in
312
several herds from 3 governorates (Salem et al., 2005). In this report, we investigated the
313
ecological, genomic and virulence characteristics of M. paratuberculosis isolates from 13
314
governorates representing different geographical and demographic regions within Egypt. PCR
315
testing of fecal samples from animals with chronic diarrhea indicated the presence of M.
316
paratuberculosis in 13.6% of the tested animals and a herd prevalence that reached 65.4%. This
317
prevalence level is considered high, especially considering that Johne’s disease in Egypt was
14
318
only recently identified as a problem for dairy herds (Salem et al., 2005). It is possible that this
319
high prevalence has resulted from the absence of any efforts to control the infection while
320
importation of dairy cows from Johne’s disease enzootic countries continues to increase. Given
321
the low sensitivity of molecular methods for detecting the presence of M. paratuberculosis
322
infection (de Juan et al., 2006), the within herd prevalence of infected animals in Egypt (range
323
12-40%) is underestimated and deserves more attention. Notably, our ecological analysis of M.
324
paratuberculosis indicates that it is widespread in both upper and lower Egypt regardless of the
325
population density – all the more reason to enact a national control program. A cornerstone for
326
such a control program would be to characterize strains causing the infection and devise
327
biosecurity measures to limit the spread of infection. In this report, we examined the genomic
328
features and diversity of M. paratuberculosis isolates and identified their virulence phenotypes.
329
To our knowledge, this is the first attempt to genotype and characterize strains of M.
330
paratuberculosis circulating in Egyptian dairy herds.
331
Following isolation of M. paratuberculosis from dairy herds, genotyping methods
332
focused on IS900 and IS1311 indicated that recent isolates from Egypt belong to the bovine
333
strain, MAP-C, as expected. Unlike sheep, which can be infected by MAP-C or MAP-S
334
genotypes (Forde et al., 2012; Ghosh et al., 2012), all cattle isolates in this study were strictly
335
MAP-C type. However, further genotyping with an SSR-based approach indicated more
336
diversity than was previously thought and provided enough justification for full genome-
337
sequencing to uncover further diversity that could not be detected by analysis of limited
338
sequence targets. As expected, whole genome sequences of E1 and E93 isolates of M.
339
paratuberculosis provided higher levels of analysis on a single nucleotide level. One of the
340
isolates,E1, had a high number of SNPs (N=432) relative to the standard reference strain of M.
15
341
paratuberculosis K10. The other isolate (E93) had only 50 SNPs compared to the reference
342
strain with only 12 SNPs common between both strains compared to the reference, another
343
indication of divergence among isolates from Egypt. Interestingly, when the mouse model was
344
used to assess virulence of each isolate, we were able to identify significant differences in
345
parameters of virulence for each strain suggesting the relative attenuation of the E1 strain which
346
had the highest number of SNPs. It is possible that such attenuation could be related to the
347
divergence from the standard strain.
348
In mice, mycobacterial colonization levels of the E93 isolate, which had fewer SNPs,
349
were similar to the levels of infection with M. paratuberculosis JTC1285, a recent field isolate
350
from cattle sequenced by our group (Hsu et al., 2011). In addition, histological analysis of
351
murine tissue indicated another divergence of the frequency and distribution of lesions induced
352
by each of the examined strains. While the E1 strain induced less granulomatous reaction in
353
livers, it provoked more lymphocytic recruitment in the intestine as indicated by enlarged
354
Peyer’s patches. On the contrary, the E93 strain induced more granulomatous reaction in livers
355
with much less enlargement of Peyer’s patches, most likely because of its ability to survive in
356
mesenteric lymph nodes compared to the E1 strain (Fig. 5). It is possible that more virulent
357
strains of M. paratuberculosis induce stronger responses with more lesions in livers and spleens
358
while dampening these responses in lymphatic organs (Peyer’s patches) to increase their
359
persistence. More attention needs to be given towards understand mechanisms behind this and
360
other potential strategies used by M. paratuberculosis to prolong its persistence in the host. All
361
of these observations suggest that M. paratuberculosis is undergoing adaptive changes to fit the
362
Egyptian environment, assuming that the M. paratuberculosis infection is recently introduced
363
with dairy cow importation from Europe and USA. Given the recent identification of M.
16
364
paratuberculosis in dairy herds in Egypt (Salem et al., 2005), recent importation is more likely
365
than a scenario where continuous evolution of M. paratuberculosis within herds has occurred for
366
long-periods while the disease went unnoticed. However, with the observed strain divergence on
367
the nucleotide levels we would expect to observe a different profile of Johne’s disease in Egypt,
368
as confirmed by some of field observations (e.g. reduced milk yield with little or no diarrhea).
369
Despite the worldwide prevalence of Johne’s disease, with documented presence in
370
Europe (Nielsen & Toft, 2009), Australia (Eppleston et al., 2014) and the USA (Lombard,
371
2011b), little attention is paid to this disease in the Middle East. In an earlier report examining
372
camels from Saudi Arabia (Ghosh et al., 2012), isolates of M. paratuberculosis belonging to the
373
MAP-S group were identified, indicating different sources of infection in the two countries
374
despite their geographical relatedness. Both the current study in Egypt and that in Saudi Arabia
375
demonstrate the power of genotyping to track the source of infection and postulate a scenario for
376
disease transmission. With the known trend of importing cattle herds from the USA and Europe
377
to Egypt, it is not surprising to see that the isolates identified in this study were of the bovine
378
type, closely related to the cattle isolates from the USA that we previously analyzed (Hsu et al.,
379
2011). However, analysis of more isolates from different regions in the Middle East could
380
improve the epizoological studies, at least on the regional level. They could also decipher the
381
dynamics of Johne’s disease transmission among herds raised in neighboring regions. Overall,
382
our approach of limited genotyping of all isolates followed by targeted whole genome
383
sequencing of divergent isolates has provided us a better understanding of virulence and
384
pathogenesis of M. paratuberculosis. Such an approach could be applied to examining Johne’s
385
disease in other countries or examining similar chronic infection in dairy herds.
386
17
387
Acknowledgement.
388
The authors would like to thank Sarah A. Marcus, Hazem F. M. Abdelaal and Aubrey Berry for
389
reading the manuscript and for constructive suggestions. We would also like to acknowledge the
390
help of Dr. Sarah Paige for generating the disease map in Fig. 1. Experiments in this report were
391
supported by USDA-NIFA #2013-67015-21347 to AMT and a grant from US-Egypt Scientific
392
board # 1937 to both AMT and AA.
393
18
394
References
395 396 397 398
Amonsin, A., Li, L. L., Zhang, O., Bannantine, J. P., Motiwala, A. S., Sreevatsan, S. & Kapur, V. (2004). Multilocus short sequence repeat sequencing approach for differentiating among Mycobacterium avium subsp paratuberculosis strains. Journal of Clinical Microbiology 42, 1694-1702.
399 400 401 402
Bannantine, J., Wu, C. w., Hsu, C. & other authors (2012). Genome sequencing of ovine isolates of Mycobacterium avium subspecies paratuberculosis offers insights into host association. BMC Genomics 13, 89.
403 404 405
Barrett, D. J., Good, M., Hayes, M. & More, S. J. (2006). The economic impact of Johne's disease in an Irish dairy herd: A case study. Irish Veterinary Journal 59, 282-+.
406 407 408
Castellanos, E., de Juan, L., Dominguez, L. & Aranaz, A. (2012). Progress in molecular typing of Mycobacterium avium subspecies paratuberculosis. Res Vet Sci 92, 169-179.
409 410 411 412
Cetinkaya, B., Egan, K., Harbour, D. A. & Morgan, K. L. (1996). An abattoir-based study of the prevalence of subclinical Johne's disease in adult cattle in south west England. EpidemiolInfect 116, 373-379.
413 414
Collins, M. T. (1996). Diagnosis of paratuberculosis. VetClinNAm 12, 357-371.
415 416 417 418
Collins, M. T., Wells, S. J., Petrini, K. R., Collins, J. E., Schultz, R. D. & Whitlock, R. H. (2005). Evaluation of five antibody detection tests for diagnosis of bovine paratuberculosis. Clin Diagn Lab Immunol 12, 685-692.
419 420 421
Darling, A. C., Mau, B., Blattner, F. R. & Perna, N. T. (2004). Mauve: multiple alignment of conserved genomic sequence with rearrangements. GenomeRes 14, 1394-1403.
422 423 424 425 426
de Juan, L., Alvarez, J., Romero, B., Bezos, J., Castellanos, E., Aranaz, A., Mateos, A. & Dominguez, L. (2006). Comparison of four different culture media for isolation and growth of type II and type I/III Mycobacterium avium subsp paratuberculosis strains isolated from cattle and goats. ApplEnvironMicrobiol 72, 5927-5932.
427 428 429 430
Eppleston, J., Begg, D. J., Dhand, N. K., Watt, B. & Whittington, R. J. (2014). Environmental Survival of Mycobacterium avium subsp. paratuberculosis in Different Climatic Zones of Eastern Australia. Appl Envir Microbiol 80, 2337-2342.
431
19
432 433
ESRI (2011).ESRI: ArcGIS Desktop: : Environmental Systems Research Institute: Redlands, CA.
434 435 436 437 438
Forde, T., Kutz, S., De, B. J., Warren, A., Ruckstuhl, K., Pybus, M. & Orsel, K. (2012). Occurrence, diagnosis, and strain typing of Mycobacterium avium subspecies paratuberculosis infection in Rocky Mountain bighorn sheep (Ovis canadensis canadensis) in southwestern Alberta. JWildlDis 48, 1-11.
439 440 441 442
Ghosh, P., Hsu, C., Alyamani, E. J. & other authors (2012). Genome-wide analysis of the emerging infection with Mycobacterium avium subspecies paratuberculosis in the Arabian camels (Camelus dromedarius). PLoS One 7, e31947.
443 444 445 446
Ghosh, P., Wu, C. W. & Talaat, A. M. (2013). The Intracellular Life of Mycobacterium avium subsp. paratuberculosis during Macrophage Stress: A Key Role for the Alternative Sigma Factor, SigH. Infect Immun.
447 448 449 450
Hines, M. E., Stabel, J. R., Sweeney, R. W. & other authors (2007). Experimental challenge models for Johne's disease: A review and proposed international guidelines. VetMicrobiol 21, 197-222.
451 452 453 454
Hsu, C.-Y., Wu, C.-W. & Talaat, A. M. (2011). Genome-wide sequence variations among Mycobacterium avium subspecies paratuberculosis isolates: A better understanding of Johne's disease transmission dynamics. Front Microbiol 2, 236.
455 456 457 458
Kasnitz, N., Köhler, H., Weigoldt, M., Gerlach, G. F. & Möbius, P. Stability of genotyping target sequences of Mycobacterium avium subsp. paratuberculosis upon cultivation on different media, in vitro- and in vivo passage, and natural infection. Veterinary Microbiology.
459 460 461 462
Li, L., Bannantine, J. P., Zhang, Q., Amonsin, A., May, B. J., Alt, D., Banerji, N., Kanjilal, S. & Kapur, V. (2005). The complete genome sequence of Mycobacterium avium subspecies paratuberculosis. ProcNatlAcadSciUSA 102, 12344-12349.
463 464 465
Lombard, J. E. (2011a). Epidemiology and economics of paratuberculosis. VetClinNorth AmFood Anim Pract 27, 525-535, v.
466 467 468
Lombard, J. E. (2011b). Epidemiology and Economics of Paratuberculosis. Veterinary Clinics of North America: Food Animal Practice 27, 525-535.
469
20
470 471 472 473
Lombard, J. E., Gardner, I. A., Jafarzadeh, S. R., Fossler, C. P., Harris, B., Capsel, R. T., Wagner, B. A. & Johnson, W. O. (2013). Herd-level prevalence of Mycobacterium avium subsp. paratuberculosis infection in United States dairy herds in 2007. Prev Vet Med 108, 234238.
474 475 476 477 478
Moss, M. T., Sanderson, J. D., Tizard, M. L., Hermon-Taylor, J., El Zaatari, F. A., Markesich, D. C. & Graham, D. Y. (1992). Polymerase chain reaction detection of Mycobacterium paratuberculosis and Mycobacterium avium subsp silvaticum in long term cultures from Crohn's disease and control tissues. Gut 33, 1209-1213.
479 480 481
Nielsen, S. r. S. & Toft, N. (2009). A review of prevalences of paratuberculosis in farmed animals in Europe. Preventive Veterinary Medicine 88, 1-14.
482 483 484 485 486
Perna, N. T., Mayhew, G. F., Posfai, G., Elliott, S., Donnenberg, M. S., Kaper, J. B. & Blattner, F. R. (1998). MOLECULAR EVOLUTION OF A PATHOGENICITY ISLAND FROM ENTEROHEMORRHAGIC ESCHERICHIA COLI O157-H7. Infection & Immunity 66, 3810-3817.
487 488 489 490
Pradhan, A. K., Mitchell, R. M., Kramer, A. J. & other authors (2011). Molecular Epidemiology of Mycobacterium avium subsp. paratuberculosis in a Longitudinal Study of Three Dairy Herds. Journal of Clinical Microbiology 49, 893-901.
491 492 493 494
Raizman, E. A., Fetrow, J. P. & Wells, S. J. (2009). Loss of income from cows shedding Mycobacterium avium subspecies paratuberculosis prior to calving compared with cows not shedding the organism on two Minnesota dairy farms. J Dairy Sci 92, 4929-4936.
495 496 497
Salem, M., Zeid, A. A., Hassan, D., El-Sayed, A. & Zschoeck, M. (2005). Studies on Johne's disease in Egyptian cattle. J Vet Med B Infect Dis Vet Public Health 52, 134-137.
498 499 500 501
Shin, S. J., Wu, C.-W., Steinberg, H. & Talaat, A. M. (2006). Identification of Novel Virulence Determinants in Mycobacterium paratuberculosis by Screening a Library of Insertional Mutants. InfecImmun 7, 3825-3833.
502 503 504 505
Shin, S. J., Han, J. H., Manning, E. J. B. & Collins, M. T. (2007). Rapid and reliable method for quantification of Mycobacterium paratuberculosis using the BACTEC MGIT 960 system. Journal of Clinical Microbiology, JCM.
506 507 508 509
Sonawane, G. G. & Tripathi, B. N. (2013). Comparison of a quantitative real-time polymerase chain reaction (qPCR) with conventional PCR, bacterial culture and ELISA for detection of subsp. infection in sheep showing pathology of Johne's disease. Springerplus 2, 45. 21
510 511 512 513 514
Stabel, J. R., Bosworth, T. L., Kirkbride, T. A., Forde, R. L. & Whitlock, R. H. (2004). A simple, rapid, and effective method for the extraction of Mycobacterium paratuberculosis DNA from fecal samples for polymerase chain reaction. Journal of Veterinary Diagnostic Investigation 16, 22-30.
515 516 517 518
Stevenson, K., Hughes, V. M., de Juan, L., Inglis, N. F., Wright, F. & Sharp, J. M. (2002). Molecular characterization of pigmented and nonpigmented isolates of Mycobacterium avium subsp paratuberculosis. Journal of Clinical Microbiology 40, 1798-1804.
519 520 521 522
Talaat, A. M., Reimschuessel, R. & Trucksis, M. (1997). Identification of mycobacteria infecting fish to the species level using polymerase chain reaction and restriction enzyme analysis. VetMicrobiol 58, 229-237.
523 524 525 526
Tamura, K., Dudley, J., Nei, M. & Kumar, S. (2007). MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) Software Version 4.0. Molecular Biology and Evolution 24, 15961599.
527 528 529
Tamura, K., Stecher, G., Peterson, D., Filipski, A. & Kumar, S. (2013). MEGA6: Molecular Evolutionary Genetics Analysis version 6.0. Mol Biol Evol 30, 2725-2729.
530 531 532 533
Wells, S. J., Whitlock, R. H. & Lindeman, C. J. (2002). Evaluation of bacteriologic culture of pooled fecal samples for detection of Mycobacterium paratuberculosis. American Journal of Veterinary Research 63, 1207-1211.
534 535 536 537 538
Wells, S. J., Collins, M. T., Faaberg, K. S., Wees, C., Tavornpanich, S., Petrini, K. R., Collins, J. E., Cernicchiaro, N. & Whitlock, R. H. (2006). Evaluation of a Rapid Fecal PCR Test for Detection of Mycobacterium avium subsp. paratuberculosis in Dairy Cattle. Clinical and Vaccine Immunology 13, 1125-1130.
539 540 541 542 543
Whittington, R., Marsh, I., Choy, E. & Cousins, D. (1998). Polymorphisms in IS1311, an insertion sequence common to Mycobacterium avium and M. aviumsubsp.paratuberculosis, can be used to distinguish between and within these species. Molecular and Cellular Probes 12, 349358.
544 545 546 547
Whittington, R. J. & Sergeant, E. S. G. (2001). Progress towards understanding the spread, detection and control of Mycobacterium avium subsp paratuberculosis in animal populations. Aust vet j 79, 267-278.
548
22
549 550 551
Wu, C.-W., Glasner, J., Collins, M. T., Naser, S. & Talaat, A. M. (2006). Whole Genome Plasticity among Mycobacterium avium subspecies: Insights from Comparative Genomic Hybridizations. JBacteriol 188, 711-723.
552 553 554 555
Wu, C. W., Schmoller, S. K., Shin, S. J. & Talaat, A. M. (2007). Defining the stressome of Mycobacterium avium subspecies paratuberculosis in vitro and in natrually infected cows. JBacteriol 189, 7877-7886.
556 557 558 559
Wynne, J. W., Seemann, T., Bulach, D. M., Coutts, S. A., Talaat, A. M. & Michalski, W. P. (2010). Resequencing the Mycobacterium avium subsp. paratuberculosis K10 genome: improved annotation and revised genome sequence. JBacteriol 192, 6319-6320.
560 561 562
23
563
Tables.
564 565
Table 1: Results of molecular (PCR) and bacteriological analyses of the fecal samples Herds
566 567 568 569
Governorate
No. of Animals
Alexandria
550
4
No. of positive herds 3
Al-Behira Al-Menoufia
440 120
4 2
Al-Daqahlia
100
Al-Sharqia
Samples No.
PCR Positive
Culture Positive
360
42
2
4 0
198 45
60 0
2 0
2
1
67
5
1
220
2
0
30
0
0
Kafr El-Sheikh
180
2
0
23
0
0
Al-Ismailia
110
1
0
29
0
0
Al-Fayoum Asiut Al-Giza Bani Suef Al-Gharbia Dumiat Total
45 50 414 280 42 25 2576
1 1 3 2 1 1 26
1 1 3 2 1 1 17 (65.4%)
36 20 381 60 32 9 1290
2 3 43 18 2 1 176 (13.6%)
1 0 0 0 1 1 8
No. of Herds
Table 2. Comparison of fecal PCR vs. ELISA testing of suspected cases of Johne's disease Pos. PCR
570 571 572 573 574
Fecal samples
Neg. PCR
Total
Pos. ELISA
14*
1
15
Neg. ELISA Total
2** 16
39 40
41 56
*Culture positive in 2 cases **Culture positive in 1 case
24
575 576 577
Table 3. Summary of reference assembly whole genome sequencing of M. paratuberculosis E1 and E93 isolates. Strain
E1
E93
Reference
K-10
K-10
Reference length
4,832,589
4,832,589
Consensus length
4,780,944
4,786,089
% Homology
98.93
99.04
Average coverage
653.06
782.48
Zero coverage counts
346
216
Read length
~101bp
~101bp
Read counts
35,189,942 44,696,724
All mapped reads
32,336,792 39,713,546
578 579
25
580
Table 4. A list of selected SNPs in coding sequences of M. paratuberculosis E1 and E93. E1 allele
E93 allele
JTC1285 allele
SNPs in other sequenced strains*
N/S**
Gene
Function
K-10 allele
MAPK0386
Conserved hypothetical protein
T
G
G
T
Yes
N
MceA2
Mce family protein
G
G
A
G
No
N
MceA2
G
G
C
G
No
S
G
C
C
G
Yes
S
MAPK1064
Mce family protein Glycosyl transferases hemolysin-like protein
A
C
C
A
Yes
N
MAPK1125
Conserved hypothetical protein
G
G
A
G
Yes
S
MAPK1125
Conserved hypothetical protein
T
G
T
T
Yes
S
MAPK1234
Arabinose efflux permease
A
A
C
A
Yes
N
MAPK1687
Conserved hypothetical protein
T
C
T
C
Yes
S
GlnE
Glutamine synthase
T
C
C
T
Yes
N
Pks12
Polyketide synthase
G
T
T
G
Yes
N
MAPK2263
A
G
A
A
No
N
MAPK2083
PPE repeat protein 2-polyprenyl-6methoxyphenol hydroxylase
CCC
CCC
CTT
CCC
No
N
MAPK2261
PE family protein
T
C
T
T
No
N
MAPK2275
Conserved hypothetical protein
A
G
G
G
Yes
S
MAPK2285
3-ketosteroid-delta1-dehydrogenase
T
A
A
A
Yes
S
CydD
ATP-binding protein, transporter
G
T
T
G
Yes
N
MAPK2641
Flavoprotein
C
T
T
C
Yes
N
SpeE
Spermidine synthase
G
A
G
A
Yes
S
MAPK2850
Trypsin-like serine proteases
C
T
C
C
Yes
S
MAPK3655
Mce family protein
C
T
C
C
No
N
MAPK3712
Conserved hypothetical protein
A
C
C
A
Yes
N
MAPK0603
26
581 582 583 584
MAPK3840
Conserved hypothetical protein
A
G
G
A
Yes
S
MAPK3933
Fucose permease
A
G
G
A
Yes
N
SNPs in other strains* refers to other sequenced isolates reference before (Hsu et al., 2011); N/S** refers to N for Non-synonymous single nucleotide polymorphism and S for Synonymous SNPs.
27
585 586
Figure Legends.
587
Fig. 1. Ecology of M. paratuberculosis in Egypt. The prevalence of Johne’s disease on the
588
animal level ranged between 5 to 30.3% in 9 governorates (4 governorates were negative).
589
Animal infection rate/governorate is color-coded based on the key to the left of the map. All
590
results were from fecal PCR analysis performed on 2576 Holstein cows in 13 governorates in
591
Egypt. Map generated using ArcGIS software (ESRI, 2011).
592 593
Fig. 2. Strain typing of M. paratuberculosis isolates from Egypt. A) Ethidium bromide stained
594
3% agarose gel of PCR amplicons of the IS1311 gene from different Egypt isolates, following
595
restriction digestion with Hinf1. For each set, both undigested (first lane) and digested products
596
(second lane) are shown (Left). Control reactions from M. paratuberculosis K10 (type C) and
597
M. paratuberculosis 397 (Type S) are also included (Right). B) A phylogenetic tree inferred
598
using the Neighbor-Joining method based on short sequence repeat (SSR) typing of 10 loci
599
(Amonsin et al., 2004). The bootstrap percentages (1000 replicates) are shown next to the
600
branches. The tree is drawn to scale and the evolutionary distances were computed using the
601
Maximum Composite Likelihood method in MEGA 6 (Tamura et al., 2013).
602 603
Fig.3. A circular genome map of newly identified single nucleotide polymorphism (SNPs)
604
in M. paratuberculosis isolates from Egypt. The outer circle shows the genomic scale. The
605
second circle shows the location of the 4,395 ORFs in the M. paratuberculosis K-10 genome.
606
Genes (olive green) on the forward strand are shown outside of the baseline; genes (light green)
607
on the reverse strand are shown inside of the baseline. Inner circles show all E93 associated
608
SNPs (magenta) and all E1 associated SNPs (blue). The figure was generated with GenVision
609
software (DNAstar, Madison, WI).
610
28
611
Fig 4. Whole genome analysis of single nucleotide polymorphisms (SNPs) of M.
612
paratuberculosis isolates from Egypt. A) Number of SNPs in M. paratuberculosis E1 and E93
613
while compared with M. paratuberculosis K-10 reference genome. A total of 50 and 432 SNPs
614
were associated with E93 and E1, respectively. The numbers of transition, transversion and
615
intergenic SNPs were also calculated. B) A phylogenetic tree inferred using the Neighbor-
616
Joining method based on the predicted SNPs of each strain compared with M. paratuberculosis
617
K-10 reference genome. The numbers at each branch are the bootstrap s and branches
618
corresponding to partitions reproduced in less than 50% bootstrap replicates are collapsed. The
619
tree is drawn to scale and the evolutionary distances were computed using the Maximum
620
Composite Likelihood method using MEGA 6 (Tamura et al., 2013).
621 622
Fig. 5. Tissue colonization of M. paratuberculosis E1 and E93. The average CFU/g of tissue
623
(liver, intestine, spleen and mesenteric lymph node, MLN) at each time point is shown in
624
comparison to infection with M. paratuberculosis JTC1285 and K-10 (Data for spleen and MLN
625
from JTC1285 and K-10 are not available). Asterisk (*) indicates statistical significance
626
(p