rioninOp:?cRc

Table 1 1. Antimicrobial resistance patterns and resistance gene profiles of antimicrobia1‑resistantEscherichia coli isolates from wild animals.

Resistance gene profile

swain"o.

Site(Sample"o.)

Hostspecies Resishncepauern("IC'a _{let bla}

Rausu town, Hokkaido Pref.

44r'OL'N, L45"I)'E F.cb.I5 toMal.. ), 2OO9

r'LLbLic/p).ivutellorest around villages.

Sikil dccr.

ShimOkIIa PenInSuIal AOmO‑･i Pref.

4I･25･〜･ IJOL･5O,E

DeC･ 21･ 2008 1O DeC･ 29, 200I)

ROad･ Vl‑‑agC･^and IubIiCI PrlVIIle^‑･OreSt 'lnd trail.

JllPaneSC mnC･211‑eS.

'

Hokkaido

ttL‑

Honshu

Gifu Pref.

35Do8'‑36Q27'N, 136ot6'‑l3T39'E Jun. I,2008toDec. l3.2009

Public/ private forest, trail,road, urban ^area, and national forest ^reserve,

Wild boars, ^Japanese macaques, sika deer and ^{Japanese serows.}

Yakushima island, Kagoshima Pref.

30ol5'‑30D23'N, )30o23'‑l30o38'E Aug. 9to 18,2008,Nov. 13 to20,2009 Nationa) forestreserve, road and trail.

Japanese macaques and sika deer.

Figure 6. Locations and descriptions of research sites.

The locations offour research sues h this study ^are shown in dark gray. The latitude and longitude, sampling ^areas, sampling per.od and wildlife species causing conflicts ^or other type of interactions with llulnanS ^Or domestic animals ^are mentioned in boxes for each

research site.

50

A

B

Figure 7. Random amp]ir]ed polymorpLlic I)NA

(RAPD)

profiles ofa subset of EschericJua co[L' isola(es from sika deer

(CervLLS

_rlt'PPOn)^{in Rausu town, Hokkaido}

Prefecture, Japan.

(A) ^Ml3‑RAPD profiles ^off. c()Illisolates from ^deer h ^Rausu. (B) ^DAF4‑RAPD profiles ^{ofLE. c()/i} iso)ates from deer ^{in Rausu. Lanes} 1, l2, and 22 ^‑ 100 bp DNA Ladder (TOYOt30, Osaka, Japan);

Lanes 2 to 10 ⁼ susceptible ^Er c()/I‑isolates (Strain ^No. HOK‑40. 45, 63, 7l, 87 to 9I, respectively);

Lane I 1^‑ OTC‑resistant ^E. (.a//‑isolates (HOK‑92): ^{Lanes 13 to 2] ‑} susceptib)e E‑ co[i iso)ares (HOK‑93, 95, 96, 99. too, 107. )08, ll I,and ^]20, respectively).Strains HOK‑90, 91, and 93 (Lanes 9, 10, and 13) ^{were iso)ated from} the ^same fecal sanlPle (Sample ^No. 22) carrying OTC‑resistant strain. HOK‑92. Strains ROK‑87 and 88 (Lanes 6 and 7)^{were isolated from} the sample ^{No, 20,}

Dissimilarity

O.7 0.6 ^{O.5 0.4 O.3 0.2 0.I O.O}

St‑rain^No. Sample No.

Ser. Gropup HOK‑1 ²⁴

HOK‑33 HOK‑151

16O lHOK‑99

HOK‑163 HOK‑24

‑HOK‑ l2 HOE‑ 147 HOK‑ ¹⁴⁴ HOK‑12

HOK‑10 HOK‑92

HOK‑179 H()K‑l2

HOK‑91 i‑

HOK‑2

HOK‑93 ^i‑

31* OUT

l1 OUT

38* 0126

40* 015

23* OUT

D B2 D A

41 9 1*

37

#13332;**;

OUT ^B1

10 4*

45*

44

3Z*

^{our B1}

3

43* 0‑uT 43* OUT

27* OUT

26* 0124 22* OUT

23* 0 20*

12 22* OUT

A BI BI BI BI B1

A

.."...4.?..."....".."..."..."...

?.2* OUT 1* 0146

k‑

^{Our B2}

22* OUT B2

HOK‑45 18 OUT B1

Figure 8. Dendrogram of Eid7eric12ia colt isolates from sika deer

(Cer1'ilSni'PPOll)

ⁱⁿ

Raus‑u ^town based on random amplirled polymorphic DNA

(RAPD)

prorlles.

Dissimilarity among isolates is based on combined M1 3‑ and DAF4‑RAPD profiles. Dissmilarity scale is^attop left.Vertical dashed ,lineindie‑ates the c‑utofffor ^a c‑luster(Dice's_cog/mcient ^< 0.2).

Cl‑usters^are shown in light gray. An OTC‑resistarlt E. coli isolates is writlen in block letters̲

Susceptible E. colt isolates from the‑fecal sample Carrying OTC‑resistant isolate^are annotated with ^a(.

Fo‑ur pairs off. coif isolates lViih identical ^RAPD profTlles^are annotated with #l‑4.Genetic diversit5t, off. tiOli^was obse‑rved in sarnples annotated vyith^*. Abbreviations are as follows: Ser^‑ serotype, OUT ^‑ 0‑serotype tlntyPable, Gro‑up ^‑ ph}7logenetie‑ group.

52

A

1.5OO I.OOO

Figure 9. Random amplified polymorphic I)NA

(RAPD)

profHes of ^a subset of

EschericJu'u call. isolates from wild animals in Shimokita Peninsula, Aomori PrefectuI.e, Japan.

(A) ^{M I3‑RAPD} prortles ofE. c''/iisolates from"ild animals in Shimokita. (B) DAF4‑RAPD profiles off. c()/iisolates from wild animals in Shimokha. ^{Lanes I,} l2, and ^{23 ‑} too bp DNA Ladder

(TOYOBO, Osaka, Japan); ^{Lanes 2 to 1I,} _{13, and} ^{14 ‑} susceptible ^E, co/J'iso)ates from Japanese macaques(strainNo. ^{SH7M‑)05, IO6, 110, 118. 122,} 132, 138, ^{14l. 149,} 153. 159,and ^160,

respectively);^{Lane 15 to 18 ‑} muLtidrug‑resistant E, cn/t'isolates from ^{a raccoon} dog (SHTM‑166 ^{to 169,} respectively);^{Lanes 19 to 22 ‑} susceptible E. c(,//Aiso)ates from ^a bear (SHJM‑178 ^to ]81, respectively).

I)ksiLlliFaThr

llLJl/I

^J

TV I

I.

Sam PTL.

No. rlost Scrr Group

SrllM‑L32 S[ lTM‑SS SHrM‑]49 SrlIM‑I 82 S[‑ILM‑l Zt6 SII IM‑202 S[ lIM‑2O7 S1‑uM‑246

SfLH M‑252 S]1IM̲28()

S] l]M‑Sol SJ.nM‑327

SILllM‑334 SI IIM‑35LI S[ TrM‑g4

8̲i tS8 9O 97 I)8 LH )()()

lll alS )22 I2ti 1:10 I 3(I 38

)I

EIM‑3t;3 HM‑2ti3.〜

HIM‑166

lil Ill Hl

fit

nl

i)J Ill )II TIE lil lu tll l1[

TH l[E ).n Hl Sl‑I SI‑I STY Sl]

SLI Sl1 SIT STl

* M;ICil(1uC OtJr [3=

l{dCLLOOn d()i E(aecoon (log i h.rotv

li5.(I"r,rL.r r.hm..unt I [7q ltaL‑LLL>On doll

̲̲..I).i?"L'.uDIHr.llhb..a.Tt.I)).I""01(1(I...J)

77 RiluO(utILtg OLTT ri2

M‑T4 M‑I) M‑6b

MT.9.O."

H:i.qi

M‑lob M‑ltd M‑llti M‑l22 M‑I.l4 M‑25t;

M‑262 M‑=7O M‑3lJ M‑323 A.t7:I.:..a

Macnque OLJrr Li=

i3;‑MAu'mLt‑‑.

^{.‑0& ‑.I).2J.}

51)

FM IM JM IM IM lh,I lM IM

i5

7() 80 82 I5q 160 198 S[1lM‑2TO SHIM‑214 SLITM‑2I9 S111M‑226 SHJM‑1JiS Sl‑[LM‑241 Sl1[M‑2bb SLJTM‑29O Sl uM‑342 SHIM‑34f}

Sl lEM‑350

i?HIRi̲:̲;i.8

SI SI

,a:u^Sl

MitCu()LLL1 E7.

1h' 27

=l) :10 85 85 qJ q5

Ill 10 Mitt.ilquL, I

I(I.a.

M.,LL,aq"C

JIM‑26 i,2 5+ [lare ()

.::.::i:I.

I ^..

::...

..:.I.

:.

Figure IO. Dendrogram ofEscJ,eTt'ChL'O CO[t'isolates ^{from wild animaJS in} Shimokita Peninsula based ^on radom amplified po]ymorphic I)NA (RAPD) _prorI]eS,

Dissinli)arit), among isolates js based on combined M I3‑ and I)^F4‑RAPD profiles. Dissi.Tl'llarity scale is shown attop felt vertical dashed line indka(es the cuton'fbr ^a cluster (Dice's_coemeient ^< 0.2). CltJSterS are Shown in light gray,

Multidrug‑resistant E. co/L' isolates ^{are wr‑Etten} h block letters. SLISCePtible E. (.a/i ;solate5 From ^{a raccoon} dog sample ^are annotated with ^**̲ Four groups orE. co/)Iisolates From Japanese macaques (Mac(7C.ujlLS'.ulu)with identical R^PD prorlles

are annotated with # I‑4. E. co[L' isolates from Japanese serows (C(TPricor77I'.".ri.ypl(S)and Japanese hares (LepuJV bl.(LCJlyllrll.Y)

are annotated with iLI and i.2.respectively. Genetic diversiL.v orE. i.0/i^was observed in si"TIPles annotated wi(h ^*.

Abbrevhtions are as Follows: Ser. ^‑ selOtyPe, OUT ^‑ 0‑serotype untypable. Group ^‑ phy[ogerLetic groLIP.

54

CHAPTER 3.

Antimicrobial Resistance and Genetic Diversity of Commensal Escherichia coli from Feces of Wild Animals and Livestock in the Republic of Zambia

INTRODUCTION

In the Republic of Zambia

(Zambia),

^{an inland} country in central Africa, wildlife conservation is very important ecologically and economically. Zambia is endowed with ^a high

diversity

of wildlife

(S14),

^and nature‑oriented tourism is a

potential

major

^{industry to earn foreign exchange and reduce}

^poverty (S4, S14).

^More

than 30% of the country's land area is conserved ^as l9 national parks and 35 Game Management Areas

(GMAs) (S14, S16).

^{GMAs are buffer zones} surrounding the national parks

(S16).

^Only non‑consumptive ^use of natural ^resources

(e.g.tourism)

^is

allowed in the national parks, whereas consumptive ^use

(e.g.

^{hunting and harvesting}

wild

plants),

^{residence, and economical activities are also allowed in GMAs}

(S14).

Some residents in GMAs are employed ^as guides ^{or scouts} for the tourism and the

wildlife management

(S

^13,

75).

Agriculture is also ^a

major

^{industry in Zambia}

(S4).

^7l.6% of the working population

(>̲

¹² years

old)

^{is engaged} in agriculture

(S5),

^{largely by} subsistence farming in rural ^areas

(S4).

^Cattle and goats ^are the most and the second ^most important livestock, respectively

(2,62).

In rural ^area, they are traditionally managed ^on natural pastures

(2,62).

^{In GMAs,} especially where local people have pastoral transhumance culture, livestock mingle with wild animals when they graze ^on natural pastures

(74).

^In

such mingling there isa high risk of pathogen transmission between livestock and wild animals

(73).

^{For example, bovine} tuberculosis and brucellosis have been persistently reported in Kafue lechwe

(Kobus

^leche

kafuensis),

^{which are endemic antelopes in}

Zambia and mingle with cattle grazing ^on the flood plain ofKafue Flats

(71,

^{72, 74,}

85).

Both diseases are thought ^to have been originally transmitted from cattle ^to lechwe

antelopes

(63,72),

^{but now they circulate in both lechwe and} cattle populations

(72,

^74,

85).

^Bovine tuberculosis is presumably infected through a respiratory ^route caused by

56

close intra‑ and interspecies interactions

(25, 74).

Brucellosis is more likely to be transmitted through environmental contamination

(7l).

^Therefore, it is necessary ^to monitor the close interactions and bacterial transmission between livestock and wild animals.

The results of Chapter 2 imply that host species and animal behaviors affect the transmission of commensal E. coli in wild animals. Cattle and wild antelopes such

as lechwe are taxonomically close species, both belonging to

Cetartiodactyla,

^Bovidae

(4, 43).

^{There are} similarities between them in behaviors such ^as grazing and their gregarious ^natures

(4,

^43,

7l).

^Therefore, it is possible that commensal E. coli may be transmitted between livestock and wild antelopes, and that commensal E. coli may be

used ^{as an} indicator to monitor the close interactions and bacterial transmission between them. Performing research similar ^to that done in Japan and comparing the results between the ^two countries will also be useful ^to elucidate the factors affecting the

spread of antimicrobial‑resistant E. coli in wild animals.

To determine how much the livestock/wildlife interactions affect the

antimicrobial resistance and transmission of commensal E. coli in animals, the following areas in Zambia were selected ^as research sites: South Luangwa National Park in Eastern Province; Lochinvar National Park in Southem Province; Chaminuka

game ^{reserve near} Lusaka, the capital

city

^{of Zambia; and two local farms. South}

Luangwa National Park is the most preferred tourist destination in Zambia

(S14).

^{In the}

nearest GMA from the sampling points of this study, livestock‑raising isnot so popular compared ^to other districts in Eastern Province

(S4).

In Lochinvar National Park and the surrounding ^area, cattle mingle with lechwe antelopes ^on the flood plain

(74).

^{This leads}

to bovine tuberculosishrucellosis transmission between lechwe antelopes and cattle ^as

kept in the environment close ^to their natural habitat. Chaminuka staff also keep goats

as their livestock ^on the premises. Cattle at ^one of the ^two farms

(Farm I)

usually graze in the nood plain ^near Lochinvar National Park. Cattle at the other farm

(Farm 2)

^{in a}

small ^town named Monze did _not have _contact with wild animals in any of the research

sites.

The

objective

^{of this chapter was to elucidate the factors affecting the}

antimicrobial resistance and transmission of commensal E. coli in wild animals and livestock. The antimicrobial resistance and genetic characteristics of commensal E. coli

from wild animals and livestock were compared between the areas with livestock/wildlife interactions

(Lochinvar

^{National Park, Chaminuka, and Farm}

1)

^and

those without the interactions

(South

^{Luangwa National Park} and Farm

2).

MATERIALS ANI) METIIODS Research sites and sample collectioyn

Feces of wild animals and livestock were collected in South Luangwa National Park

(12o27'‑13o36'S, 31oOO'‑32oO2'E),

^{Lochinvar National Park}

(l5o43'‑16oO1'S,27o11'‑27o19'E),

^Chaminuka

(15oO9'S,28o29'E),

and ^two local farms

(Farms

1 and

2).

^{The locations} of each research site and sampling points ^are shown in Figures ll ^to l3. In order ^to compare environmental E. coli isolates with those from

animals, soils and/or ^{water were} also collected in the ^two national parks and Chaiminuka. Feces of wild animals ^were collected in South Luangwa National Park on

August 2009 add August 2010

(Fig.l2).

^{Feces of wild animals and cattle were collected}

in Lochinvar National Park and the ^two farms on August 2010

(Figs.

1l and

13).

^Feces

of captive antelopes and goats ^were collected in Chaminuka on August 2009. The captive antelopes in Chaminuka had _not been treated with antimicrobials according ^to

58

Chaminuka staff. Permission for sample collection in the present study ^was acquired from the authorities concerned.

Wild animals ^were searched for by car and feces on the ground ^were collected.

Soil and ^{water were} also collected during the drive. Collected feces and soil ^were stored in plastic bags ^{at room} temperature until ^use. Water was stored in sterilized plastic bottles _at room temperature until ^use.

Isolation and identirICation ofEscherichia coli

Collected samples ^were usually transported ^to

University

of Zambia

(UNZA)

for bacteri'al isolation. In South Luangwa,‑ bacterial isolation was performed ^on site because it took ^more than ^two days for transportation. Feces and soil ^were diluted to

approximately 5%

[w/v]

^{or 10%}

[w/v]

ⁱⁿ sterilized PBS. Bacterial isolation from fecal/soil suspensions and ^{water were} performed ^as described in Chapter 2. Colonies or

bacterial plaques ^on DHL agar plates ^were picked up and stabbed into soft nutrient

agars

(Nissui,

^Tokyo,

Japan)

in microtubes for transportation ^to Gifu University, Japan.

At Gifu University, all bacteria ^were streaked ^onto DHL agar plates. Up to four lactose‑fermenting colonies ^were selected for each sample. The identification and stock

ofE. coli ^were performed ^as described in Chapter 2.

Screening for antimicrobial resistance ofE. coli

The antimicrobial

susceptibility

^{of all E.} coli isolates ^was screened by disc difAISion method with SN disc

(Nissui)

^following manufacturer's instructions

(Table 12).

Escherichia coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853 were

used ^as

quality

^standard.

Broth microdilution method

(49)

^{was performed on} E. coli isolates that ^were regarded

as "resistant" by disc difbsion method according ^to CLSI

(Table 12).

^{A subset of}

described in Chapter 2.

Detection of antimicrobial resistance genes

PCR for tetracycline‑resistance

(tet)

genes

(78)

^was performed ^on oxytetracycline

(OTC)‑resistant

^E. coli isolates and ^a subset of susceptible isolates as

described in Chapter 2

(Table l3).

PCR for streptomycin

(SM)‑resistance

genes

(aadA,

strA,

strB)(12,32)

^was

performed ^on SM‑resistant E. coli isolates and ^a subset of susceptible isolates. Primers used ^to amplify each gene ^were listed in Table 13. Each PCR mixture

(25 p1)

^was

prepared using TaKaRa Ex TaqTM

(Takara)

^following manufacturer's instructions. PCR for aadA and strB ^was performed under the conditions described elsewhere

(12, 32).

PCR for strA ^was performed under the following conditions: initial denaturation for 5

min ^at 95oC, 40 cycles of I min ^at 95oC, 30 s at 55oC, 30 s at 72oC, and ^a final extension step of 5 min ^at 72oC. Electrophoresis and visualization of PCR amplicons

were performed ^as described in Chapter 2.

Typing phylogenetic groups ofE. colt

Escherichia coli isolates ^were classified into four phylogenetic groups, A, B 1_, B2 and D, by multiplex PCR

(21)

^{as described in Chapter 2}

(Table 13).

Random amplified polymorphic DNA

(RAPD)

analysis

Rand.m amplifled p.lym.,phic DNA

(MPD)

analysis ^was perf.rmed

;.

^E.

coli isolates &om herbivores

(1‑4

isolates per

sample)

^{as described in Chapter 2}

(Table 13).

Statistical analysis

The prevalence of antimicrobial‑resistant E. coli isolates was compared between wild animals and livestock by Fisher's exact test. The prevalence of

antimicrobial‑resistant E. coli in wild animals/livestock ^was compared among research

60

sites by Fisher's ^{exact test.} Statistical analysis ^was performed ^as described in Chapter 2.

RESULTS Number of samples and the prevalence ofE. coh'

A total of 168 fecal samples ^were collected from wild animals in two years ^at three research sites

(South

^{Luangwa, Lochinvar} and Chaminuka; Table

14).

^{The animal}

species sampled included 10 species of herbivores, four species of omnivores, and three species of camivores. A total of 306 E. coli isolates ^were recovered from 90 fecal samples

(53.6%) (Table 14).

^A total of 39 fecal samples ^were collected from domestic goats and cattle in ^two years ^at the three research sites‑

(Chaminuka,

^Farms 1 and 2;

Table

15). Twenty‑four

E. coli isolates ^were recovered from 12 fecal samples

(30.8%) (Table 15).

Environmental samples included 18 soil samples in South Luangwa, two soil samples in Chaminuka, and ^{one water} sample from each of South Luangwa, Lochinvar and Chaminuka.

Twenty‑two

E. coli isolates ^were recovered from six soil

samples

(26.1%)

collected in South Luangwa. No E. coli ^was isolated from _water samples.

Antimicrobial resistance ofE. colt from wild animals and livestock

Two E. coli isolates from ^two lion samples in South Luangwa were resistant

to OTC

(Tables

^{14 and}

16).

^One E. coli isolate from ^a goat sample in Chaminuka was

resistant ^to SM and OTC

(Tables

¹⁵ and

16).

^{The total} prevalence of antimicrobial‑resistant E. coli ^was 2.22%

(2/90)

^{in wild animals and 8.33%}

(1/12)

ⁱⁿ

livestock

(Tables

14 and

15).

^The diffTerence was not significant between wild animals and livestock

(Fisher's

^{exact test, P‑0.32,}

n‑102).

^{The prevalence of}

antimicrobial‑resistant E. coli in wild animals ^at each research site ^was 2.86%

(2/70)

ⁱⁿ

difference was not significant among research sites

(Fisher's

^{exact test, P‑1,}

n‑90).

^The

prevalence of antimicrobia1‑resistant E. coli in livestock ^ateach research site^was 20.0%

(1/5)

in Chaminuka and O% in Farm 1

(0/5)

and Farm 2

(0/2)(Table15).

^{The difference}

was not significant among research sites

(Fisher's

^{exact test, P‑1,}

n‑12).

^No

antimicrobial‑resistant E. coli ^was isolated from soil.

A goat‑derived E. coli isolate resistant ^to SM and OTC in Chaminuka

(Strain

No.

CMK‑58)

carried the strA, strB, and

tet(B)

^genes

(Table 16).

OTC‑resistant E. coli isolates from lions in South Luangwa were not examined for tet genes by PCR

(Table 16).

^No resistance genes ^were detected in the susceptible E. coli tested.

Genetic relatedness ofE. coli from wild animals and livestock

RAPD analysis ^was performed ^on 101 E. coli isolates from wild herbivores

and 14 isolates from livestock. Zero to nine reproducible fragments were amplified by M13‑RAPD analysis

(Fig.14‑A),

^{and zero to} 13 reproducible fragments were amplified by DAF4‑RAPD analysis

(Fig.14‑B).

A total of 106

types

of combined RAPD profiles

were observed and 87.0% _ofE. coli isolates

(100/115)

^{had a unique MPD}

type (Fig.

15).

^Sixteen clusters

(clusters

^{1 to} 16; Fig. 15, light

gray)

^{were observed, but 59.1% of}

E. coli

(68/l15)

^{did not belong to} any clusters.

The antimicrobial‑resistant E. coli isolates ^were differentiated from E. coli isolates susceptible ^to all antimicrobials examined in this study. The goat‑derived E.

coli isolate resistant ^to SM and OTC

(CMK‑58;

Fig. 15, block

letters)

and ^a susceptible isolate from the ^same fecal sample

(CMK‑59;

^{Fig. 15,}

#)

^had unique RAPD profiles unrelated ^to each other.

The

genotypes

^{of E. coli isolates were compared between different host}

species, between individuals of each host species, and within ^an individual fTecal sample.

Comparing E. coli isolates between host species, ^none of the isolates had RAPD

62

profiles identical to those of isolates from different host species. Seven clusters

(clusters

3, 6, 7, 8, 9, 10 and 15; Fig. 15, light

gray)

^includedE. coli isolates from ^more than one

animal species belonging to the order Cetartiodacty1a. Three clusters

(clusters

^{6, 8, and}

10; Fig. l5, light

gray)

^{included isolates from different animal species in different}

research sites. In ^a comparison of E. coli isolates between individuals of each host species, the predominant E. coli strains in hces seemed ^to be more specific ^to individuals than ^to animal species. Seven clusters

(clusters

l, 4, 5, 8, 12, 13, and 16; Fig.

15, light

gray)

^included E. coli isolates from different fecal samples of the ^same animal species. Only one pair of isolates from different impala

(Aepyceros melampus)

^samples

(SL‑221

and 225; Fig. l5, cluster

1)

^{had identical RAPD} profiles and belonged to the

same phylogenetic group

(Group B1).

^The genetic diversity _{of E. coli within} an

individual fecal sample ^was observed in impalas, waterbucks

(Kobus ellipsl'PTymnuS),

lechwes, pukus

(Kobus va71donii),

^kudus

(Tragelaphus strepsiceros),

^{a giraffe}

(GiraHa camelopa1.dalis),

^{an African elephant}

(Loxodonta africana),

^{and a goat}

(Fig.

^{14, lanes 9}

and 10; Fig. 15,

*).

DISCUSSION

Two E. coli isolates from lions in South Luangwa were resistant ^to OTC

(Table 16).

^One E. coli isolate from ^a goat in Chaminuka was resistant ^to SM and OTC, carrying the strA, strB, and

let(B)

^genes

(Table 16).

^The strA, strB, and

let(B)

^{genes are}

widely distributed plasmid‑borne genes

(14,

^18,

94).

^Itwas reported that strA, strB and

tet(B)

^{genes are often linked with other} antimicrobial resistance genes ^or heavy metal resistance genes

(18, 94).

^{Both research sites}

(South

^Luangwa and

Chaminuka)

^were

little affected by the antimicrobial ^use in veterinary medicine. Therefore, the resistance

horizontal gene transfer ^or cross‑selection for other antimicrobials/metals without the

use of SM and OTC. Further analyses of the ^tet and the str genes detected in this study such ^as plasmid pro filing ^or

conjugative

^{assay may be useful to understand the}

dissemination and persistence mechanism of these resistance genes in _nature.

The prevalence of antimicrobial‑resistant E. coli ^was very low in both wild animals and livestock, and there ^{was no} significant difference between them. No

significant difference was observed between research sites, regardless of the presence of livestock/wildlife interactions. The prevalence of antimicrobial‑resistant E. coli in livestock in the present study ^was equivalent ^to that in healthy cattle in l989

(6.7%;

7/105) (79).

^{In Zambia,} cattle ^are occasionally treated with antimicrobials

(62,69),

^but

usually goats do _not receive veterinary ^care

(2).

^{This low frequency} of antimicrobial ^use is probably the ^reason for the low prevalence of antimicrobial‑resistant E. coli in livestock over the past 20 years. Unless the antimicrobial ^use increases in livestock

management in Zambia, mingling of wild animals and livestock will ^not affect the prevalence of antimicrobial‑resistant E. coli in wild animals. Other epidemiological markers ^are needed ^{to use} with antimicrobial resistance ^to monitor the close interactions

or bacterial transmission between livestock and wild animals.

The results of

genotyping

^{of E. coli}

(Fig.15)

^imply that the predominant E.

coli strains from Zambian wild herbivores and livestock are more specific ^to individuals than ^to animal species. The predominant E. coli strains in the intestines of Zambian herbivores might rarely colonize in the intestines of other individuals. Otherwise,

transmitted E. coli strains might remain ^as minor components of the E. coli population.

H.wever, the E. coli in individual feces is genetically

dive;se (Fig.

^15,

'*).

_{This result}

suggests that undetectable E. coli strains may exist in E. coli populations of Zambian herbivores and be transmitted among animals. More E. coli isolates should be analyzed

64

to detect the transmission of minor strains.

The individual

specificity

^{of E.} coli and/or the genetic

diversity

of E. coli within ^an individual fecal sample ^were observed in all research sites, regardless of the presence of livestock/wildlife interactions. Some researchers reported that the genetic

diversity

ofE. coli within ^an individual was lower in livestock or captive wild animals than in free‑ranging wild animals

(30,42).

^Goldberg and his colleagues suggested that anthropogenic factors such ^as forest fragmentation affected the E. coli transmission in primates

(35,36).

^{At present,} the livestock/wildlife interaction did _not appear ^to affect the E. coli transmission ^or its genetic

diversity

^{in the} research sites of this study.

However, _not all E. coli isolates ^were analyzed by RAPD analysis, and the number ofE.

coli isolates analyzed ^was smaller in Lochinvar and Chaminuka than that in South Luangwa. Further analyses ^are needed ^to statistically compare the genetic

diversity

ofE.

coli population between research sites.

The relatively close relatedness ofE. coli ^was observed between different host species belonging to the order Cetartiodactyla

(clusters

3, 6, 7, 8, 9, lO and 15; Fig. 15, light

gray).

^{Moreover, in the case} of clusters 6, 8 and 10

(Fig.

^{15, light}

gray),

^{E. coli}

strains ^were isolated from the animals in different research sites

(South

^Luangwa and

Lochinvar).

^{South Luangwa National Park is over 500 km} away from Lochivar National Park, and GMAs around these national parks ^are also separated

(Sl4).

^{Because there is}

no route for wild animals ^{to move} between these ^two national parks, the relatively close

relatedness of E. coli is unlikely ^to be caused by direct transmission. All E. coli strains in clusters 6, 8 and 10

(Fig.

^{l5, light}

gray)

^{were isolated from ruminants or}

hippopotamuses

(Hippopotamus amphibius).

It is found that there is little difference in the digestive functions of African ruminants

(38).

^{Hippos are foregut fermenters as well}

adapt ^to the microenvironment in the intestines of ruminants ^or foregut fermenters. If so,

E. coli is^more likely to be transmitted between cattle and wild ruminants grazing ^on the

same pasture than between livestock and other wild animals. To confirm this hypothesis, the E. coli isolates from omnivores and camivores need ^to be compared with those of herbivores.

In summary, this study suggests that epidemiological markers other than

antimicrobial resistance ^are needed ^to monitor the close interactions or bacterial transmission between livestock and wild animals in Zambia. The

genotyping

^results

imply that livestock/wildlife interactions have not affected the genetic

diversity

of E.

coli within ^an individual animal ^or the E. coli transmission in wild animals ^so far.

However,

genotyping

^{results also imply that the digestive function such as foregut}

fTermentation might be one of the factors affecting E. coli colonization. Therefore, the

research should be continued ^on commensal E. coli from wild ruminants and livestock in areas with livestock/wildlife interactions.

66

S UMA4ARY

In Zambia, pathogen transmission caused by close interaction between livestock and wild antelopes isa problem in animal health and wildlife conservation. To

elucidate whether the livestock/wildlife interaction affects the antimicrobial resistance

and transmission of commensal E. coli in animals ^{or not,} E. coli from wild animals and livestock were screened for antimicrobial resistance and ^were characterized by

genotyping

^{in four areas in Zambia. A} total of 168 fecal samples ^were collected from wild animals and 39 fecal samples ^were collected from livestock. A total of 306 E. coli isolates were recovered from 90 fecal samples in wild animals and 24 E. coli isolates

were recovered from 12 fecal samples in livestock. Two E. coli isolates from ^two lions

were resistant ^to OTC. One isolate from a goat ^was resistant ^to SM and OTC, carrying the strA, strB, and

tet(B)

^genes.

^Genotyping

^ofE. coli isolates from herbivores suggests that the predominant E. coli strains vary among individuals and that the E. coli population within ^an individual animal is genetically diverse. At present, there is no

evidence that the livestock/wildlife interactions affect antimicrobial resistance, transmission ^or genetic

diversity

of commensal E. coli in wild animals and livestock.

However,

genotyping

^{results also imply that digestive functions such as foregut}

fermentation might be one of the factors affecting E. coli colonization. Therefore, further study should be performed ^on commensal E. coli &om wild ruminants and livestock in areas with livestock/wildlife interactions.

ドキュメント内 Wildlife Ecology and Environmental Monitoring using Escherichia coli as an Indicator (ページ 53-109)