Identification and genetic mapping of bovine chemokine genes
expressed in epithelial cells
Michael P. Heaton, William W. Laegreid, Craig W. Beattie,* Timothy P.L. Smith, Steven M. Kappes
U.S. Department of Agriculture, ARS, U.S. Meat Animal Research Center (MARC), State Spur 18D, P.O. Box 166, Clay Center,
Nebraska 68933-0166, USA
Received: 2 July 1998 / Accepted: 24 September 1998
Abstract. RNA fingerprinting by arbitrarily primed (RAP)-PCR
was used to identify two bovine genes that were differentially
expressed in epithelial cells during an inflammatory response.
RNA fingerprints revealed two differentially amplified transcripts
when monolayers of Madin-Darby bovine kidney (MDBK) cells
were stimulated with Escherichia coli O157:H7 lipopolysaccha-
ride (LPS) in combination with cycloheximide (CX). Sequence
analysis showed that both transcripts encoded members of the
alpha C-X-C chemokine family; one was interleukin 8 (IL-8), and
the other was a protein closely related to bovine growth-regulated
protein (GRO)-␥ (89% identical). The latter putative epithelial cell
inflammatory protein was designated ECIP-1. IL-8 and ECIP-1
genes were placed on the cattle genetic map with single-nucleotide
polymorphism (SNP) markers amplified from genomic DNA.
Multi-point linkage analysis indicated that the gene locations were
indistinguishable from those of serum albumin (ALB) and vitamin
D-binding protein (GC) on bovine Chromosome (BTA) 6. In hu-
mans, ALB and GC are located near IL-8, GRO-␥, and seven other
alpha chemokines on Chr 4 (HSA 4q11-4q13), suggesting that this
gene cluster has been conserved on BTA6. These results provide a
starting point for characterizing allelic variation in chemokine
genes and their role in the pathogenesis of bacterial infections in
Bacterial infections in cattle represent a potential risk to animal
and human health. Enterohemorrhagic Escherichia coli and Sal-
monella typhimurium DT104 from beef sources have recently
caused serious outbreaks, resulting in human morbidity and mor-
tality (Bell et al. 1994; MMWR 1997; Glynn et al. 1998). The first
lines of defense against enteric infections are the epithelial mucosa
and the immune mechanisms activated by contact with bacterial
envelopes. This contact initiates responses leading to local inflam-
mation and the eventual clearance of infection (Svanborg et al.
1996; Kagnoff and Eckmann 1997). The important role of inflam-
matory response genes in disease resistance has been underscored
by recent reports of polymorphisms in human genes that confer
either resistance or susceptibility to infection (Dean et al. 1996;
Smith et al. 1997b; Altare et al. 1998; Jong et al. 1998; Winkler et
al. 1998). A related goal in livestock is to identify polymorphisms
that genetically restrict the infection of cattle by human foodborne
pathogens. To that end, the aim of this study was to identify and
map bovine genes that are modulated in the early phase of bovine
epithelial cell exposure to E. coli O157:H7 LPS.
Materials and methods
In vitro model for stimulated bovine epithelial cells.
model consisting of MDBK cells treated with purified LPS from E. coli
O157:H7 was selected to fingerprint bovine epithelial cell expression.
MDBK cultures were utilized because a bovine intestinal epithelium cell
line is not available. LPS is an outer membrane component and virulence
factor of the Gram-negative envelope. LPS from E. coli O157:H7 is of
particular interest because it has a well-known ability to induce expression
of a variety of genes in other cell systems and has been proposed to play
a role in disease development (Taylor 1995; Bilge et al. 1996; Karpman et
al. 1997). Furthermore, a significant anti-LPS antibody response is induced
in cattle that are infected with E. coli O157:H7 (Laegreid et al. 1998).
Detection of mRNAs associated with the immediate-early transcriptional
responses to LPS was enhanced by treating some cultures with CX. This
protein synthesis inhibitor is known to stabilize some rapidly labile mRNA
species and induce transcription of others (Zuckerman et al. 1991; Mc-
Clelland et al. 1994). For LPS purification, E. coli O157:H7 (ATCC
43895) was grown in brain-heart infusion broth (Difco Laboratories, De-
troit, Mich.). Purified LPS from E. coli O157:H7 was prepared as previ-
ously described (Leive et al. 1968; Laegreid et al. 1998). MDBK cells
(ATCC CCL-22) were cultured in Minimum Essential Medium (MEM)
containing Earle’s salts, 2 m
-glutamine, 100 U penicillin G, 100 U
streptomycin sulfate (Gibco BRL Gaithersburg, Md.), and 10% fetal bo-
vine serum (Atlanta Biologicals, Norcross, Ga.). Confluent cell cultures
were divided and plated into six 9.6-cm
tissue culture treated wells at a
density of 3 × 10
cells/well in 3 ml MEM. Monolayers were grown to 95%
confluence (approximately 48 h), washed once in 1 ml MEM, and treated
with 3 ml fresh MEM plus LPS (3 g/ml). Control samples were treated
with 3 ml MEM alone. Where indicated, CX (10 g/ml, Sigma Chemical
Company, St. Louis, Mo.) was added 1 h prior to harvest.
Arbitrary 10-mer oligonucleotide primers including OP-
26-03 (5Ј-CTTTCTACCC-3Ј) and OP26-19 (5Ј-GATCATAGCC-3Ј) were
manufactured by Operon Technologies, Inc. (Alameda, Calif.). DEAD2+
(5Ј-GATGAGGCTGA-3Ј) and KINASEA1+ (5Ј-GAGGGTGCCTT-3Ј)
were kindly provided by Michael McClelland, Sidney Kimmel Cancer
Center, San Diego, Calif. Oligonucleotides for amplification of bovine
genomic IL-8 (sense, 5Ј-ACACATTCCACACCTTTCTACC-3Ј; anti-
sense, 5Ј-AGCAGACCTCGTTTCCATTGG-3Ј) and ECIP-1 (sense, 5Ј-
CCCGAAGCTCCCATGGTTAAG-3Ј; antisense 5Ј-CCCTGGAAAC-
CAGCCATTCTC-3Ј) were synthesized on a 1000
(Beckman Coulter, Inc., Fullerton, Calif.).
RNA fingerprints were obtained as previously described with
slight modification (Liang and Pardee 1992; Welsh et al. 1992; Ralph et al.
1993; McClelland et al. 1994; Wong and McClelland 1994). Briefly, RNA
was isolated from fresh tissue culture with a silica-gel-based membrane
purification system as per manufacturer’s instruction (Qiagen, Inc., Santa
Clara, Calif.). After isolation, 100 ng/l of total RNA was treated with
amplification grade deoxyribonuclease I according to the manufacturer’s
instructions (Gibco BRL). Since variations in RNA quality and concentra-
tion may introduce significant effects on amplification results, each RNA
preparation was fingerprinted three times at twofold serial dilutions. These
samples were then loaded in adjacent lanes of a single fingerprint gel.
First-strand cDNA synthesis was accomplished in a 10-l reaction
Correspondence to: M.P. Heaton
*Present address: University of Minnesota, College of Veterinary and
Pathobiology, St. Paul, Minnesota 55108, USA.
Mammalian Genome 10, 128–133 (1999).
© Springer-Verlag New York Inc. 1999