lntramammary Challenge with Escherichia coli Following Immunization with a Curli-Producing Escherichia coli' D. A. TODHUNTER, K. L. SMITH, and J. S. HOGAN Department of Dairy Science The Ohio State University Ohio Agricultural Research and Development Center Wooster 44691
L. NELSON Alfa-Laval Agrilntemational Tumba, Sweden
in milk production after challenge was greater for cows immunized with E. coZi pCRL65/A012. Immunization of dairy cattle with a curli-producing strain of E. coli did not protect against experimental intramammary challenge during lactation. (Key words: intramammary challenge, Escherichia coli immunization, curli)
ABSTRACT
Holstein and Jersey cattle were immunized with a curli-producing strain of Escherichia coli (pcRL65/A012) or a noncurli-producing strain (pU C l8 / HBlO1) to determine differences in resistance to establishment of experimental intramammary infection. Cows (n = 6 per group) were immunized at 14 d prior to drying off,7 d of involution, and at calving with 3 x 1Olo E. coli in Freund's Incomplete Adjuvant. At 30 d of lactation, one mammary quarter of each cow was infused with a wild strain of E. coli (727). Escherichia coli 727 was isolated from a naturally occurring intramammary infection and produced curli. All challenged quarters became infected, and all cows developed acute clinical mastitis. Geometric mean duration of intramammary infections was 6 d for both immunization groups. All infections were spontaneously eliminated within 10 d. No differences occurred between immunization groups in blood selenium and glutathione peroxidase activity, plasma selenium, number of E. coli 727 isolated from secretion after challenge, rectal temperature and SCC response, clinical status of mammary quarters, or DMI. Reduction
Received May 11, 1990. Accepted October 25, 1990.
1991 J Dairy Sci 74:819-825
Abbreviation key: curli - = group immunized with Escherichia coli pUCl8/HI3101, curli + = group immunized with Escherichia coli pcRL65/A012, IMI = intramammary infection, PBS = phosphate-buffered saline. INTRODUCTION
An Escherichia coli surface organelle was identified and named curli (11). Curli are thin, coiled, wiry fibers composed of a single subunit, curlin. Ability to bind to fibronectin was mediated by curli structures. Binding to fibronectin may be a virulence-associated property for colonization of wounds and fibronectincoated surfaces (11). The curlin gene, crl, was present and transcribed in most E. coli isolates; however, only certain strains were able to assemble curlin into the curli surface organelle. Approximately half of E. coli isolated from bovine intramammary infections (IMI) were able to bind soluble lZI-labeled fibronectin (11). Escherichia coli isolated from bovine IMIwere shown to lack many of the virulence factors associated with E. coli in other diseases (16). Serum resistance was the onlv factor iden-
819
820
TODHUNTER ET AL.
noncurli-producing E. coli vaccine. The objec- 727. Escherichia coli 727 was isolated initially tives of this study were to monitor the effects from a naturally occurring IMI in the Ohio of vaccination on general cow health and to Agricultural Research and Development Center determine if curli-immunized cows were more dajr herd. Escherichia coli 727 is a curli + resistant to establishment of IMI following E. strain based on fibronectin binding. Escherichia coli 727 bound 22.8% of total [1251&ibronectin. coli challenge at 30 d of lactation. Escherichia coli 727 was stored in trypticase soy broth (BBL Microbiology Systems, Becton MATERIALS AND METHODS Dickinson and Co., Cockeysville, MD) plus 20% glycerol (Mallinckrodt Inc., Paris, KY) at Animals -7o'C. Prior to infusion, E. coli 727 was Animals were Holstein (n = 9) and Jersey (n streaked for isolation on esculin blood agar and = 3) dairy cattle from the Ohio Agricultural incubated for 24 h at 37'C. A single isolated Research and Development Center, Wooster, colony was transfemd to brain-heart infusion OH. At 14 d prior to drying off, cows were broth (Difco Laboratories) and incubated at assigned randomly by breed to one of two 37'C for 12 h. Following incubation, E. coli immunization groups: immunized with E. coli 727 was diluted in filter sterilized (.2 pkf, pUC18/HB101 (curli -) and immunized with Gelman Sciences, Ann Arbor, MI) PBS to E. coli pCRL65/A012 (curli +). Six cows were achieve approximately 50 cfu/ml of PBS. Acassigned to each immunization group. Two Jer- tual number of colony-forming units infused was determined by plating an aliquot in triplisey cows were in the curli A vaccine cate onto the surface of a MacConkey agar dose was composed of 3 x 10 formalin-killed E. coli in 5 ml of phosphate-buffered saline (Becton Dickinson Microbiology Systems, (PBS) emulsified in 5 ml of Freund's Incom- Cockeysville, MD) plate. Diluted bacteria were plete Adjuvant @ifco Laboratories, Detroit, transported immediately on ice to the dairy MI). Vaccines were provided by Alfa-Laval barn and infused into mammary quarters. MamAgriIntemational, Tumba, Sweden. Curli pro- mary quarters were infused with bacteria 4 h duction by E. coli was measured by ability to postmilking. Only uninfeaed mammary quarbind soluble [1251Jfibronectinas described by ters were used for E. coli 727 challenge. Ols6n and Normark (11). Percentage of total Rectal temperature was recorded at 0, 3, 6, soluble [125flfibronectinbinding was 30% for 9, 14, 24, 48, 72, and 96 h postchallenge. The E. coli pCRL65IAO12 and less than 1%for E. clinical status of secretion was determined on coli pUC18/HBlOl. foremilk at the time of sampling and coded 1 to Cows were immunized subcutaneously on 5 as by Hogan et al. (10). Abnormal secretion the upper rib cage posterior to the scapula (5 was defined as the presence of clots, flakes, etc. d s i d e ) at 14 d prior to drying off and at Clinical status of quarters was recorded 0,3,6, calving. Cows also received an inmammary 9, 14,24,48,72,%, 168,240,480, and 720 h immunization at 7 d of involution. Intramam- after E. coli 727 challenge. Duration of clinical mary immunization was by infusion of 2.5 ml signs was the number of days the quarter was of vaccine into each mammary quarter. Rectal clinical converted to loglo. temperature, condition of the injection site, and general cow health were recorded 0, 12,24,48, Cell Counts and 72, and 96 h postimmunhtion. All mammary Somatic Colony-Forming Units In quarters of all cows were infused with a dry Secretlon Postchalienge cow antibiotic preparation (300 mg cephapirin Aseptically obtained milk samples were colbenzathine; Tomorrow, Agricultural Products, Division of Bristol-Myers Co., Dewitt,NY) on lected at 0, 3, 6, 9, 14, 24, 48, 72, 96, 168, 240, 480, and 720 h postchallenge for determination the day of drying off. of SCC and bacterial counts. Somatic cell counts were determined electronically (Coulter Bacterial Challenge Electronics, Inc., Hialeah, FL) as described by At approximately 30 d of lactation, one Smith et al. (17). Direct microscopic SCC were quarter of each cow was infused with E. coli done on samples that had large clots and flakes.
PP.
Journal of Dairy Science Vol. 74,No. 3, 1991
821
IMMUNIZATION WITH CURLI-POSITIVE ESCHERICHIA COLJ
Somatic cell counts were expressed as loglo (SCC/ml). Colony-forming units of E. coli 727 were determined by plating diluted and undiluted secretion onto the surface of a MacConkey agar plate. Secretion was diluted in sterile PBS. All dilutions were done in duplicate. The minimal detection limit was .5 cfu/ml. Colony-forming units were expressed as loglo. A quarter from which no bacteria were isolated was assigned a value of -.3010 loglo cfu/ml or loglo of -5. Duration of experimental infection was the number of days between the first and last isolation of E. coZi 727; values were transformed to loglo before analysis. Milk Production and Dry Matter Intake
hfik production and DMI were recorded
daily on all cows from 7 d prior to challenge to 7 d postchallenge. Cows were milked twice daily. Cows were housed in tie stalls, and refusals were weighed daily. Percentage of change in milk production and DMI was calculated for 7 d postchallenge. Percentage of change was calculated as: 100(a - b)/b; where a = daily milk production or DMI and b = mean milk production or DMI for 7 d prior to challenge. Whole Blood Selenlum, Plasma Selenium, and Glutathione Peroxidase Actlvlty
Selenium and glutathione peroxidase (GSHpx) activity were determined on heparinized blood obtained at 30 d of lactation prior to bacterial challenge. Plasma was obtained by centrifugation of an aliquot of whole blood at 3000 x g for 5 min at 4'C. Blood and plasma Se were determined fluorometrically (12) and expressed as micrograms per milliliter. Whole blood GSHpx activity was determined spectrophotometrically by the method of Paglia and Valentine (14). Glutathione peroxidase activity was expressed as nanomoles of NADPH oxidized per minute per milligram of hemoglobin (Hb) at 25'C. The cyanomethemoglobin method was used to determine Hb concentration (2). Statistical Analysls
Comparison of mean rectal temperature response, loglo colony-forming units per mil-
A lntramammary Dry Period
0 Subcutaneous Drying Off
w Subcutaneous
5
40.0
38.5
{
1
%
/
I 0
, 12
24
48
72
96
Hours Post Immunization Figure 1. Rectal temperature response to immunization with 3 x 10" Escherichia coli pcRL65/AO12 or E. coli pUClS/HBlOl. Cows (n = 12) were immunized subcutaneously at 14 d prior to drying off (SubcutaneousDryiog off) and at calving (Subcutaneous Calving). Cows also received an i n 1 . vimmunization at 7 d of the dry p&od (Intmnamma~~ Dry Period). Values are expressed as mean f standard error.
liliter; SCC response, duration of intramammary infection, percentage of change in milk production and DMI, whole blood GSHpx activity, and plasma and blood Se was by least squares analysis of variance (18). RESULTS
There were no differences (P > .05) in mean body temperresponse to immunization between groups at any time; therefore, body temperature data were combined from both groups. Body temperature response to subcutaneous immunization at 14 d prior to drying off was approximately S'C at 12 h postimmunization (Figure 1). There was no body temperature response to immunization at calving. htramammary immunization at 7 d of involution resulted in a mean increase in rectal temperature of 2.3'C at 12 h postimmunization. Body temperature returned to preimmunization temperature at 24 h for both subcutaneous and intramammary immunization. All injection sites appeared normal following subcutaneous immunization at 14 d prior to drying off. Following subcutaneous immunizaJ o d of Dairy Science Vol. 74, No. 3, 1991
822
TODHUNTER ET AL. A Curli Positive (pCRL65,’AOlZ)
0 Curli Negative (pUC 18,’HB 101)
A C u r l 1 Positive
0 Curli Negative ( p U C l8/HB 10 1)
( p C R L 6 5 j A O 12)
41.5
!
41.0
u
s
40.5
4
2
40.0
(u
a
g
39.5
e 39.0 0
3
6
9
14 2 4 46 72 96 168240480720
Hours Post Challenge Figure 2. Colony forming units of Escherichia coli 727 isolated from secretion following intramammary challenge ’ withacurliat 30 d of lactation. Cows were immmued positive E. coli @CRL65/A012) or a curli-negative E. coli @UC18/HB101).Values are expressed as mean f standard error of six cows.
tion at calving, all sites were normal except on one cow that had a palpable 10.16-cm (4-in) diameter swelling at the injection site. All mammary glands appeared normal following intramammary immllIlization. Mean number of colony-forming units of E. coli 727 infused into quarters at 30 d of lactation was 60 for curli - immunized cows and 67.7 cfu for curli + immunjzed cows. Range of Colony-forming units infused was 13 to 85. Escherichia coli 727 was reisolated from all challenged quarters (Figure 2). Mean loglo colony-forming units per milliliter increased to approximately 4 for both groups between 9 and 14 h. Mean loglo colony-forming units per milliliter declined from 24 to 96 h postinfusion and was 1 loglo cfu/ml or less by 96 h. Escherichia coli 727 was not reisolated after 240 h from any challenged quarter. None of the cows received antibiotic therapy after challenge. There were no differences (P > .05) in loglo colony-forming units per milliliter between immunization groups. AU cows displayed systemic signs of acute clinical mastitis by 14 h postchallenge. All cows appeared normal 48 h after challenge. Secretion from all challenged quarters displayed clinical signs by 24 h postchallenge. At 7 d postchallenge, secretion from 75% of quarJournal of Dairy Science Vol. 74, No. 3, 1991
38.5 I
0
12
24
36
46
60
72
64
96
Hours Post Challenge Figure 3. Rectal temperature response to htr8-y challenge with Eschericfiu coli 727 at 30 d of lactation. Cows were immunized with a curli-positive E. coli (pCRL65/A012) or a curli-negative E . coli (pUC18/ HB101). Valua are mean f staudard error of six cows.
ters appeared normal. Secretion from a l l quaters was normal by 30 d postchallenge. Peak rectal temperature response was observed at 14 h postchallenge (Figure 3). Maximal body temperature was 40.6’C and 41.1‘C for curli - and curli + immunized cows, respectively. Body temperature returned to prechallenge values by 24 h in both groups. There were no differences (P > .05) in mean body temperature response at any time between the two groups. Somatic cells exceeded 7 logldml between 14 and 24 h for both groups (Figure 4). Mean SCC was greater than 6 logldml up to 240 h postchallenge and returned to prechallenge levels by 480 h in all quarters. There were no differences (P > .05) in duration of MI and clinical signs, maximal colonyforming units, time of maximal colony-forming units, whole blood glutathione peroxidase and Se, or plasma Se between immunization groups (Table 1). Geometric mean duration of experimental E. coli 727 was 6 d for curli - immunized cows and 6.2 d for curli + immunized cows. Maximal loglo colony-forming units per milliliter was 4.31 for curli - immunized cows and 4.41 for curli + immunized cows. Mean time of maximal colony-forming units was 50.8
823
IMMUNIZATION WITH CURLI-POSITIVEESCHERICHIA COLl
TABLE 1. Infection and blood selenium parameters m cows immunized with aukpositive and curb-negative Escherichia coli following intramamma~ychallenge (IMI).
-
SE
X
Parameter Days duration IMI, loglo Geometric mean duration M I , d Days duration clinical s i p , loglo
.78 6.03 .66
Geometric means duration clinical signs, d Time maximal cfu, h Maximal log1Cb cfu/ml Whole blood GSHpx4 Whole blood selenium, Plasma seleniam, pghl
.07
Range .60-1.04
1.10 .09
.30-.95
4.60
120
50.80
37.90 .15 8.70
4.31 82.20
,324 .137
9 -244) 3.644.65
.042 .007
X
.79 6.16 .79 6.30 14.00 4.41 97.90 .331 .174
SE
.06
Range .69-1.23
120 .10
.60-1.23
1.30 3.10 23 9.50 .021
9-24 3.72-5.22
.a24
'Cows immunized with E. coli pCRL65/A012. k o w s immunized with E . coli pUc18iHB101. 'values are expressed as mean standard m. 4Glutathi0ne peroxidase activity expressed as aanomoles of NADPH oxidized per minute per milligram of hemoglobin
*
at 25'C.
-
and 14 for curli and curli + immunized cows, respectively. The most frequent time of observed maximal colony-forming units was 9 h postchallenge. Milk production decreased 23% in curli immunized cows and 36% in curli + immunized cows during the first 24 h after challenge (Figure 5). Reduction in milk yield was greater (P < .05) in curli + immunized cows than for curli - i m r n u cows. Milk production returned to prechallenge yields 4 to 5 d after E. coli infusion. Dry matter intake decreased 4% on the day of challenge for both groups. Dry matter intake returned to normal 1 to 2 d postchallenge F igure 6). There were no differences ( P > .05) in percentage of change in DMI between immunization groups.
al. (11) proposed that curli m a y be important for wound colonization or adherence of E. coli to fibronectincoated surfaces. Approximately half of E. coli from clinical bovine IMI and fecal isolates were able to bind soluble fibronectin. Adherence to epithelial surfaces is considered important in the pathogenesis of
A Curl1 Positive (pCRL65/AO 12)
Curl1 Negative (pUC 18/HB 10 1)
7 5 i 7.0
2
\
0 65 U
rL
DISCUSSION
D
6.0
-
M
Immunization of daj. cattle with a curli5.5 producing strain of E. coli had no effect on the establishment of experimental IMI during early lactation compared with immunization with a 0 3 6 9 14 24 48 72 96 1 6 8 2 4 0 4 8 0 7 2 0 noncurli-producing strain. Somatic cell count Hours P o s t Challenge response, body temperature response, duration of IMI, duration of clinical signs, and bacterial P i 4. Somatic cell count response to intramauumy numbers isolated from secretion postchallenge challenge with Escherichia coli 727 at 30 d of lactation. were similar between the two immunization somatic cell count was expressed as log10 scc!/ml. cows were immwith a curli-positive E. coli @cRLa5/ groups. Binding to soluble fibronectin was me- A012) or a curli-negative E. coli @UC18/HB101).Values diated by curli surface structures, and Oldn et are mean standard ermr of six cows.
*
Journal of Dairy Science Vol. 74, No. 3, 1991
824
TODHUNTER ET AL. 0 Curh Negative (pUC 18/HB 10 1)
A Curli Positive (pCRL65 / A 0 12)
2o 1
A Curli Positive a, 3
T
(pCRL65/AO 12)
20.0
0 Curli Negative (pvc IE/HB i o 1)
1
15.0 10 0 5 0 0 0 -5.0
*R.05 --45 I MEANO
r
I
7 Days
Prior
-10.0
1
2
3
4
5
6
7
Day Post Challenge
Figure 5. Change in daily milk production in cows challenged inb 'ly with Escherichia coli 727 at 30 d of lactation. Cows were immunized with a curli-positive E . coli @CRL65/A012)or a curli-negative E . coli @UC18/ HBlO1). Values are expressed as percentage of change in milk production based on the mean milk production 7 d prior to challenge. Values are mean f standard error of six cows.
mastitis due to Streptococcus agalactiae and Staphylococcus aweus (7, 13). Several studies demonstrated that E. coli does not adhere to mammary gland epithelium (6, 7, 13), and adherence would not appear to be a virulence factor for E. coli mastitis. The role of curli in adherence of E. coli to mammary gland epithelium is unknown; however, immunization with a curli-producing E. coli strain had no influence on experimental infections caused by a curliproducing wild strain of E. coli. Although all challenged quarters became infected and all cows developed acute clinical mastitis, experimental IMI were eliminated relatively quickly. Geometric mean duration of IMIwas 6 d, and the longest duration of an IMI was 10 d. Other challenge studies demonstrated the short duration of most experimental E. coli IMI (9,15). The majority of naturally occurring E. coli IMI are also of short duration (17). Although bacteria were eliminated quickly, SCC remained elevated in all challenged quarters for approximately 20 d postchallenge. The largest decrease in milk production was observed 1 d after E. coli challenge, followed Journal of Dairy Science Vol. 74. No. 3, 1991
1
MEANO 7Days Pr1or
1
2
3
4
5
6
7
Day Post Challenge
Figure 6. Change in DMI following intrvnammary challenge with Escherichia coli 727 at 30 d of lactation. Cows were immunized with a curli-positive E. coli (pCRL65/A012) or a curli-negative E . coli @UC1%/ HB101). Values are expressed as percentage of change in DMI based on the mean DMI 7 d prior to challenge. Values are mean f standard error of six cows.
by return to normal milk production 3 to 5 d after infusion. Other studies have reported return to normal milk yield in most cows within several days of experimental E. coli challenge (7, 8, 15). In the current study, milk production returned to normal after E. coli was eliminated from the gland. Cows immunized with curliproducing E. coli had a significantly greater depression in milk yield at 24 h than did cows immunized with noncurli-producing E. coli. Rainard (15) reported that drop in milk production parallels clinical severity of the disease. There were no apparent differences in clinical severity of experimental E. coli IMI between the groups. Additionally, DMI for the days following challenge was similar between the two immunization groups. Reasons for differences in milk yield reduction between the two groups are unknown. Selenium deficiency of dairy cows influences the severity and duration of experimental E. coli IMI (4). Whole blood glutathione peroxidase and whole blood and plasma Se concentrations for immunized cows indicated a similar and adequate Se status for all challenged cows. Subcutaneous immunization of killed E. coli in adjuvant resulted in a very mild transient
IMMUNIZATION WITH CZIRLI-POSI'ITVE ESCHEMCHlA COW
body temperature increase. Swelling at the injection site was observed in one cow. Previous experience with subcutaneous injection has shown that cow movement at the time of immunization would appear to be related to swelling at the injection site. Body temperature response to intramammary immunization was approximately 2'C and returned to normal by 24 h after immunization. All mammary glands appeared normal. Previous studies have demonstrated that the dry gland is relatively refractory to the effects of E. coli endotoxin (5). All cows progressed normally through the dry period to calving. Although direct infusion of bacteria into the mammary gland is one method of determining the protective value of a vaccine, it may not be the same as natural exposure. Entry of Gramnegative bacteria into the mammary gland occurs by passage of bacteria through the streak canal into the teat cistem (3). Direct deposition of E. coli into the cistern could bypass any protective mechanisms in the streak canal, eliminating the need for bacterial factors required to colonize or pass through the streak canal. Also, the number of bacteria gaining entrance into the teat or gland cistern may be very different than with naturally occurring
IMI. The current study demonstrated that the lactating mammary gland is not protected from experimental E. coli challenge following immunization with curli-producing strain of E. coli. Challenge with E. coli at 30 d of lactation resulted in clinical mastitis. All experimental IMI were self-eliiated and of short duration. ACKNOWLEDGMENTS
This study was supported in part by AlphaLaval AgriIntemational, Sweden. The authors thank Pamela Schoenberger, Sue Romig, Heidi Rennecker, Lucinda Shock, and Sandy Renner for technical assistance. Fibronectin binding data was generously supplied by Alfa-Laval AgriIntemational, Sweden. REFERENCES lcarroll, E. J., N. C. Jain, 0. W. Schalm, and J. Lasmanis. 1973. Experimentally induced coliform mastitis: inoculation of udders with serum-sensitive
825
and serum-resistant organisms. Am. J. Vet. Res. 3 4 1143. 2Crosby, W. H., J. 1. Munn, and F. W. FWh. 1954. Standardizing a method for clinical hemoglobinometry. US Amed Forces Med. J. 5:693. 3 Eberha~t,R. J., R. P. Natzke, F.H.S. Newbould, B. Nonnecke, and P. Thompson. 1979. Coliform mastitika review. J. Dairy Sci. 621. 4Erskine. R. I., R J. Eberhart, P. J. Grasso, and R W. Scholz. 1989. Induction of Escherichia coli mastitis in cows fed selcmum-deficient or selenium-supplemented diets. Am. J. Vet. Res. 502093. 5Frost, A. J., and B. E. Brooker, 1983. The effect of Eschm'chio coli endotoxin and culture filtrate on the dry bovine mammary gland. J. Comp. Pathol. 93:211. 6Frost, A. I., and A. W. Hill. 1982. Pathogenesis of experimental bovine mastitis following a small inoculum of Escherichia coli. Res. Vet. Sci. 33:105. 7Frost, A. I., D. D. Wauasi@e, and J. B. Woolcock 1977. Some factors affecting selective adherence of microorganisms m the bovine mammary gland. Infect. Immun. 15245. 8 Hill, A. W. 1981. Factors influencing the outcome of Escherichiu coli mastitis in the davy cow. Res. Vet. Sci. 31:107. 9 Hill,A. W., A. L. Shears,and K. G. Hibbitt. 1978. The elbination of serum-resistant Escherichiu coli from experimentally infected single mammary glands of healthy cows. Res. Vet. Sci. 25239. lOHogan, J. S., K. L. Smith,K. H. Hoblet, P. S. Schoenberger, D. A. Todhunter. W. D. Hueston, D. E. Pritchard, G. L. Bowman, L. E. Heider, B. L. Brockett, and H. R Conrad. 1989. Field swvey of clinical mastitis in low somatic cell count herds. J. Dairy Sci. 721547. 11 OMn A., A. Jonsson, and S. Normark. 1989. Fibronectin binding mediated by a novel class of surface organelles on Escherichio coli. Nature (Lond.) 33:652. 120lson, 0. E.,I. Palmer, and E. Cary. 1975. Modification of the official fluorometric method for selenium in plants. J. Assoc. Offic. Agric. Chem. 58:117. 13Opdebeeck, J. P.,A. J. Frost, and D. O'Boyle. 1988. Adhesion of Stuphylococcus uureus and Escherichiu coli to bovine udder epithelial cells. Vet. Microbiol. 16:77. 14Paglia, D. E., and W. N. Valentine. 1%7. Studies on the quantitative and qualitative characterizations of erythrocyte glutathione peroxidase. J. Lab. Clin. Med. 70158. 15 Rainard, P. 1983. Experimental mastitis with Escberichio coli: kinetics of bacteriostatic and bactericidal activities. Ann. Rech. Vet. 141. 16Sanchez-Carlo, V., J. S. McDonald, and R. A. Packer. 1984. Virulence factors of Escherichia coli isolated from cows with acute mastitis. Am. J. Vet. Res. 45: 1775.
17 Smith, K.L.,D. A. Todhunter, and P.S. Schoenberger. 1985. Environmental mastitis: cause, prevalence, prevention. J. Dairy Sci. 68:1531. 18Sokal, R. R., and F. J. Rohlf. 1969. Biometry. The principles and practice of statistics in biological research. W. H. Freeman and Company, San Prancisco, CA. Journal of Dairy Science Vol. 74, No. 3, 1991
L. NELSON Alfa-Laval Agrilntemational Tumba, Sweden
in milk production after challenge was greater for cows immunized with E. coZi pCRL65/A012. Immunization of dairy cattle with a curli-producing strain of E. coli did not protect against experimental intramammary challenge during lactation. (Key words: intramammary challenge, Escherichia coli immunization, curli)
ABSTRACT
Holstein and Jersey cattle were immunized with a curli-producing strain of Escherichia coli (pcRL65/A012) or a noncurli-producing strain (pU C l8 / HBlO1) to determine differences in resistance to establishment of experimental intramammary infection. Cows (n = 6 per group) were immunized at 14 d prior to drying off,7 d of involution, and at calving with 3 x 1Olo E. coli in Freund's Incomplete Adjuvant. At 30 d of lactation, one mammary quarter of each cow was infused with a wild strain of E. coli (727). Escherichia coli 727 was isolated from a naturally occurring intramammary infection and produced curli. All challenged quarters became infected, and all cows developed acute clinical mastitis. Geometric mean duration of intramammary infections was 6 d for both immunization groups. All infections were spontaneously eliminated within 10 d. No differences occurred between immunization groups in blood selenium and glutathione peroxidase activity, plasma selenium, number of E. coli 727 isolated from secretion after challenge, rectal temperature and SCC response, clinical status of mammary quarters, or DMI. Reduction
Received May 11, 1990. Accepted October 25, 1990.
1991 J Dairy Sci 74:819-825
Abbreviation key: curli - = group immunized with Escherichia coli pUCl8/HI3101, curli + = group immunized with Escherichia coli pcRL65/A012, IMI = intramammary infection, PBS = phosphate-buffered saline. INTRODUCTION
An Escherichia coli surface organelle was identified and named curli (11). Curli are thin, coiled, wiry fibers composed of a single subunit, curlin. Ability to bind to fibronectin was mediated by curli structures. Binding to fibronectin may be a virulence-associated property for colonization of wounds and fibronectincoated surfaces (11). The curlin gene, crl, was present and transcribed in most E. coli isolates; however, only certain strains were able to assemble curlin into the curli surface organelle. Approximately half of E. coli isolated from bovine intramammary infections (IMI) were able to bind soluble lZI-labeled fibronectin (11). Escherichia coli isolated from bovine IMIwere shown to lack many of the virulence factors associated with E. coli in other diseases (16). Serum resistance was the onlv factor iden-
819
820
TODHUNTER ET AL.
noncurli-producing E. coli vaccine. The objec- 727. Escherichia coli 727 was isolated initially tives of this study were to monitor the effects from a naturally occurring IMI in the Ohio of vaccination on general cow health and to Agricultural Research and Development Center determine if curli-immunized cows were more dajr herd. Escherichia coli 727 is a curli + resistant to establishment of IMI following E. strain based on fibronectin binding. Escherichia coli 727 bound 22.8% of total [1251&ibronectin. coli challenge at 30 d of lactation. Escherichia coli 727 was stored in trypticase soy broth (BBL Microbiology Systems, Becton MATERIALS AND METHODS Dickinson and Co., Cockeysville, MD) plus 20% glycerol (Mallinckrodt Inc., Paris, KY) at Animals -7o'C. Prior to infusion, E. coli 727 was Animals were Holstein (n = 9) and Jersey (n streaked for isolation on esculin blood agar and = 3) dairy cattle from the Ohio Agricultural incubated for 24 h at 37'C. A single isolated Research and Development Center, Wooster, colony was transfemd to brain-heart infusion OH. At 14 d prior to drying off, cows were broth (Difco Laboratories) and incubated at assigned randomly by breed to one of two 37'C for 12 h. Following incubation, E. coli immunization groups: immunized with E. coli 727 was diluted in filter sterilized (.2 pkf, pUC18/HB101 (curli -) and immunized with Gelman Sciences, Ann Arbor, MI) PBS to E. coli pCRL65/A012 (curli +). Six cows were achieve approximately 50 cfu/ml of PBS. Acassigned to each immunization group. Two Jer- tual number of colony-forming units infused was determined by plating an aliquot in triplisey cows were in the curli A vaccine cate onto the surface of a MacConkey agar dose was composed of 3 x 10 formalin-killed E. coli in 5 ml of phosphate-buffered saline (Becton Dickinson Microbiology Systems, (PBS) emulsified in 5 ml of Freund's Incom- Cockeysville, MD) plate. Diluted bacteria were plete Adjuvant @ifco Laboratories, Detroit, transported immediately on ice to the dairy MI). Vaccines were provided by Alfa-Laval barn and infused into mammary quarters. MamAgriIntemational, Tumba, Sweden. Curli pro- mary quarters were infused with bacteria 4 h duction by E. coli was measured by ability to postmilking. Only uninfeaed mammary quarbind soluble [1251Jfibronectinas described by ters were used for E. coli 727 challenge. Ols6n and Normark (11). Percentage of total Rectal temperature was recorded at 0, 3, 6, soluble [125flfibronectinbinding was 30% for 9, 14, 24, 48, 72, and 96 h postchallenge. The E. coli pCRL65IAO12 and less than 1%for E. clinical status of secretion was determined on coli pUC18/HBlOl. foremilk at the time of sampling and coded 1 to Cows were immunized subcutaneously on 5 as by Hogan et al. (10). Abnormal secretion the upper rib cage posterior to the scapula (5 was defined as the presence of clots, flakes, etc. d s i d e ) at 14 d prior to drying off and at Clinical status of quarters was recorded 0,3,6, calving. Cows also received an inmammary 9, 14,24,48,72,%, 168,240,480, and 720 h immunization at 7 d of involution. Intramam- after E. coli 727 challenge. Duration of clinical mary immunization was by infusion of 2.5 ml signs was the number of days the quarter was of vaccine into each mammary quarter. Rectal clinical converted to loglo. temperature, condition of the injection site, and general cow health were recorded 0, 12,24,48, Cell Counts and 72, and 96 h postimmunhtion. All mammary Somatic Colony-Forming Units In quarters of all cows were infused with a dry Secretlon Postchalienge cow antibiotic preparation (300 mg cephapirin Aseptically obtained milk samples were colbenzathine; Tomorrow, Agricultural Products, Division of Bristol-Myers Co., Dewitt,NY) on lected at 0, 3, 6, 9, 14, 24, 48, 72, 96, 168, 240, 480, and 720 h postchallenge for determination the day of drying off. of SCC and bacterial counts. Somatic cell counts were determined electronically (Coulter Bacterial Challenge Electronics, Inc., Hialeah, FL) as described by At approximately 30 d of lactation, one Smith et al. (17). Direct microscopic SCC were quarter of each cow was infused with E. coli done on samples that had large clots and flakes.
PP.
Journal of Dairy Science Vol. 74,No. 3, 1991
821
IMMUNIZATION WITH CURLI-POSITIVE ESCHERICHIA COLJ
Somatic cell counts were expressed as loglo (SCC/ml). Colony-forming units of E. coli 727 were determined by plating diluted and undiluted secretion onto the surface of a MacConkey agar plate. Secretion was diluted in sterile PBS. All dilutions were done in duplicate. The minimal detection limit was .5 cfu/ml. Colony-forming units were expressed as loglo. A quarter from which no bacteria were isolated was assigned a value of -.3010 loglo cfu/ml or loglo of -5. Duration of experimental infection was the number of days between the first and last isolation of E. coZi 727; values were transformed to loglo before analysis. Milk Production and Dry Matter Intake
hfik production and DMI were recorded
daily on all cows from 7 d prior to challenge to 7 d postchallenge. Cows were milked twice daily. Cows were housed in tie stalls, and refusals were weighed daily. Percentage of change in milk production and DMI was calculated for 7 d postchallenge. Percentage of change was calculated as: 100(a - b)/b; where a = daily milk production or DMI and b = mean milk production or DMI for 7 d prior to challenge. Whole Blood Selenlum, Plasma Selenium, and Glutathione Peroxidase Actlvlty
Selenium and glutathione peroxidase (GSHpx) activity were determined on heparinized blood obtained at 30 d of lactation prior to bacterial challenge. Plasma was obtained by centrifugation of an aliquot of whole blood at 3000 x g for 5 min at 4'C. Blood and plasma Se were determined fluorometrically (12) and expressed as micrograms per milliliter. Whole blood GSHpx activity was determined spectrophotometrically by the method of Paglia and Valentine (14). Glutathione peroxidase activity was expressed as nanomoles of NADPH oxidized per minute per milligram of hemoglobin (Hb) at 25'C. The cyanomethemoglobin method was used to determine Hb concentration (2). Statistical Analysls
Comparison of mean rectal temperature response, loglo colony-forming units per mil-
A lntramammary Dry Period
0 Subcutaneous Drying Off
w Subcutaneous
5
40.0
38.5
{
1
%
/
I 0
, 12
24
48
72
96
Hours Post Immunization Figure 1. Rectal temperature response to immunization with 3 x 10" Escherichia coli pcRL65/AO12 or E. coli pUClS/HBlOl. Cows (n = 12) were immunized subcutaneously at 14 d prior to drying off (SubcutaneousDryiog off) and at calving (Subcutaneous Calving). Cows also received an i n 1 . vimmunization at 7 d of the dry p&od (Intmnamma~~ Dry Period). Values are expressed as mean f standard error.
liliter; SCC response, duration of intramammary infection, percentage of change in milk production and DMI, whole blood GSHpx activity, and plasma and blood Se was by least squares analysis of variance (18). RESULTS
There were no differences (P > .05) in mean body temperresponse to immunization between groups at any time; therefore, body temperature data were combined from both groups. Body temperature response to subcutaneous immunization at 14 d prior to drying off was approximately S'C at 12 h postimmunization (Figure 1). There was no body temperature response to immunization at calving. htramammary immunization at 7 d of involution resulted in a mean increase in rectal temperature of 2.3'C at 12 h postimmunization. Body temperature returned to preimmunization temperature at 24 h for both subcutaneous and intramammary immunization. All injection sites appeared normal following subcutaneous immunization at 14 d prior to drying off. Following subcutaneous immunizaJ o d of Dairy Science Vol. 74, No. 3, 1991
822
TODHUNTER ET AL. A Curli Positive (pCRL65,’AOlZ)
0 Curli Negative (pUC 18,’HB 101)
A C u r l 1 Positive
0 Curli Negative ( p U C l8/HB 10 1)
( p C R L 6 5 j A O 12)
41.5
!
41.0
u
s
40.5
4
2
40.0
(u
a
g
39.5
e 39.0 0
3
6
9
14 2 4 46 72 96 168240480720
Hours Post Challenge Figure 2. Colony forming units of Escherichia coli 727 isolated from secretion following intramammary challenge ’ withacurliat 30 d of lactation. Cows were immmued positive E. coli @CRL65/A012) or a curli-negative E. coli @UC18/HB101).Values are expressed as mean f standard error of six cows.
tion at calving, all sites were normal except on one cow that had a palpable 10.16-cm (4-in) diameter swelling at the injection site. All mammary glands appeared normal following intramammary immllIlization. Mean number of colony-forming units of E. coli 727 infused into quarters at 30 d of lactation was 60 for curli - immunized cows and 67.7 cfu for curli + immunjzed cows. Range of Colony-forming units infused was 13 to 85. Escherichia coli 727 was reisolated from all challenged quarters (Figure 2). Mean loglo colony-forming units per milliliter increased to approximately 4 for both groups between 9 and 14 h. Mean loglo colony-forming units per milliliter declined from 24 to 96 h postinfusion and was 1 loglo cfu/ml or less by 96 h. Escherichia coli 727 was not reisolated after 240 h from any challenged quarter. None of the cows received antibiotic therapy after challenge. There were no differences (P > .05) in loglo colony-forming units per milliliter between immunization groups. AU cows displayed systemic signs of acute clinical mastitis by 14 h postchallenge. All cows appeared normal 48 h after challenge. Secretion from all challenged quarters displayed clinical signs by 24 h postchallenge. At 7 d postchallenge, secretion from 75% of quarJournal of Dairy Science Vol. 74, No. 3, 1991
38.5 I
0
12
24
36
46
60
72
64
96
Hours Post Challenge Figure 3. Rectal temperature response to htr8-y challenge with Eschericfiu coli 727 at 30 d of lactation. Cows were immunized with a curli-positive E. coli (pCRL65/A012) or a curli-negative E . coli (pUC18/ HB101). Valua are mean f staudard error of six cows.
ters appeared normal. Secretion from a l l quaters was normal by 30 d postchallenge. Peak rectal temperature response was observed at 14 h postchallenge (Figure 3). Maximal body temperature was 40.6’C and 41.1‘C for curli - and curli + immunized cows, respectively. Body temperature returned to prechallenge values by 24 h in both groups. There were no differences (P > .05) in mean body temperature response at any time between the two groups. Somatic cells exceeded 7 logldml between 14 and 24 h for both groups (Figure 4). Mean SCC was greater than 6 logldml up to 240 h postchallenge and returned to prechallenge levels by 480 h in all quarters. There were no differences (P > .05) in duration of MI and clinical signs, maximal colonyforming units, time of maximal colony-forming units, whole blood glutathione peroxidase and Se, or plasma Se between immunization groups (Table 1). Geometric mean duration of experimental E. coli 727 was 6 d for curli - immunized cows and 6.2 d for curli + immunized cows. Maximal loglo colony-forming units per milliliter was 4.31 for curli - immunized cows and 4.41 for curli + immunized cows. Mean time of maximal colony-forming units was 50.8
823
IMMUNIZATION WITH CURLI-POSITIVEESCHERICHIA COLl
TABLE 1. Infection and blood selenium parameters m cows immunized with aukpositive and curb-negative Escherichia coli following intramamma~ychallenge (IMI).
-
SE
X
Parameter Days duration IMI, loglo Geometric mean duration M I , d Days duration clinical s i p , loglo
.78 6.03 .66
Geometric means duration clinical signs, d Time maximal cfu, h Maximal log1Cb cfu/ml Whole blood GSHpx4 Whole blood selenium, Plasma seleniam, pghl
.07
Range .60-1.04
1.10 .09
.30-.95
4.60
120
50.80
37.90 .15 8.70
4.31 82.20
,324 .137
9 -244) 3.644.65
.042 .007
X
.79 6.16 .79 6.30 14.00 4.41 97.90 .331 .174
SE
.06
Range .69-1.23
120 .10
.60-1.23
1.30 3.10 23 9.50 .021
9-24 3.72-5.22
.a24
'Cows immunized with E. coli pCRL65/A012. k o w s immunized with E . coli pUc18iHB101. 'values are expressed as mean standard m. 4Glutathi0ne peroxidase activity expressed as aanomoles of NADPH oxidized per minute per milligram of hemoglobin
*
at 25'C.
-
and 14 for curli and curli + immunized cows, respectively. The most frequent time of observed maximal colony-forming units was 9 h postchallenge. Milk production decreased 23% in curli immunized cows and 36% in curli + immunized cows during the first 24 h after challenge (Figure 5). Reduction in milk yield was greater (P < .05) in curli + immunized cows than for curli - i m r n u cows. Milk production returned to prechallenge yields 4 to 5 d after E. coli infusion. Dry matter intake decreased 4% on the day of challenge for both groups. Dry matter intake returned to normal 1 to 2 d postchallenge F igure 6). There were no differences ( P > .05) in percentage of change in DMI between immunization groups.
al. (11) proposed that curli m a y be important for wound colonization or adherence of E. coli to fibronectincoated surfaces. Approximately half of E. coli from clinical bovine IMI and fecal isolates were able to bind soluble fibronectin. Adherence to epithelial surfaces is considered important in the pathogenesis of
A Curl1 Positive (pCRL65/AO 12)
Curl1 Negative (pUC 18/HB 10 1)
7 5 i 7.0
2
\
0 65 U
rL
DISCUSSION
D
6.0
-
M
Immunization of daj. cattle with a curli5.5 producing strain of E. coli had no effect on the establishment of experimental IMI during early lactation compared with immunization with a 0 3 6 9 14 24 48 72 96 1 6 8 2 4 0 4 8 0 7 2 0 noncurli-producing strain. Somatic cell count Hours P o s t Challenge response, body temperature response, duration of IMI, duration of clinical signs, and bacterial P i 4. Somatic cell count response to intramauumy numbers isolated from secretion postchallenge challenge with Escherichia coli 727 at 30 d of lactation. were similar between the two immunization somatic cell count was expressed as log10 scc!/ml. cows were immwith a curli-positive E. coli @cRLa5/ groups. Binding to soluble fibronectin was me- A012) or a curli-negative E. coli @UC18/HB101).Values diated by curli surface structures, and Oldn et are mean standard ermr of six cows.
*
Journal of Dairy Science Vol. 74, No. 3, 1991
824
TODHUNTER ET AL. 0 Curh Negative (pUC 18/HB 10 1)
A Curli Positive (pCRL65 / A 0 12)
2o 1
A Curli Positive a, 3
T
(pCRL65/AO 12)
20.0
0 Curli Negative (pvc IE/HB i o 1)
1
15.0 10 0 5 0 0 0 -5.0
*R.05 --45 I MEANO
r
I
7 Days
Prior
-10.0
1
2
3
4
5
6
7
Day Post Challenge
Figure 5. Change in daily milk production in cows challenged inb 'ly with Escherichia coli 727 at 30 d of lactation. Cows were immunized with a curli-positive E . coli @CRL65/A012)or a curli-negative E . coli @UC18/ HBlO1). Values are expressed as percentage of change in milk production based on the mean milk production 7 d prior to challenge. Values are mean f standard error of six cows.
mastitis due to Streptococcus agalactiae and Staphylococcus aweus (7, 13). Several studies demonstrated that E. coli does not adhere to mammary gland epithelium (6, 7, 13), and adherence would not appear to be a virulence factor for E. coli mastitis. The role of curli in adherence of E. coli to mammary gland epithelium is unknown; however, immunization with a curli-producing E. coli strain had no influence on experimental infections caused by a curliproducing wild strain of E. coli. Although all challenged quarters became infected and all cows developed acute clinical mastitis, experimental IMI were eliminated relatively quickly. Geometric mean duration of IMIwas 6 d, and the longest duration of an IMI was 10 d. Other challenge studies demonstrated the short duration of most experimental E. coli IMI (9,15). The majority of naturally occurring E. coli IMI are also of short duration (17). Although bacteria were eliminated quickly, SCC remained elevated in all challenged quarters for approximately 20 d postchallenge. The largest decrease in milk production was observed 1 d after E. coli challenge, followed Journal of Dairy Science Vol. 74. No. 3, 1991
1
MEANO 7Days Pr1or
1
2
3
4
5
6
7
Day Post Challenge
Figure 6. Change in DMI following intrvnammary challenge with Escherichia coli 727 at 30 d of lactation. Cows were immunized with a curli-positive E. coli (pCRL65/A012) or a curli-negative E . coli @UC1%/ HB101). Values are expressed as percentage of change in DMI based on the mean DMI 7 d prior to challenge. Values are mean f standard error of six cows.
by return to normal milk production 3 to 5 d after infusion. Other studies have reported return to normal milk yield in most cows within several days of experimental E. coli challenge (7, 8, 15). In the current study, milk production returned to normal after E. coli was eliminated from the gland. Cows immunized with curliproducing E. coli had a significantly greater depression in milk yield at 24 h than did cows immunized with noncurli-producing E. coli. Rainard (15) reported that drop in milk production parallels clinical severity of the disease. There were no apparent differences in clinical severity of experimental E. coli IMI between the groups. Additionally, DMI for the days following challenge was similar between the two immunization groups. Reasons for differences in milk yield reduction between the two groups are unknown. Selenium deficiency of dairy cows influences the severity and duration of experimental E. coli IMI (4). Whole blood glutathione peroxidase and whole blood and plasma Se concentrations for immunized cows indicated a similar and adequate Se status for all challenged cows. Subcutaneous immunization of killed E. coli in adjuvant resulted in a very mild transient
IMMUNIZATION WITH CZIRLI-POSI'ITVE ESCHEMCHlA COW
body temperature increase. Swelling at the injection site was observed in one cow. Previous experience with subcutaneous injection has shown that cow movement at the time of immunization would appear to be related to swelling at the injection site. Body temperature response to intramammary immunization was approximately 2'C and returned to normal by 24 h after immunization. All mammary glands appeared normal. Previous studies have demonstrated that the dry gland is relatively refractory to the effects of E. coli endotoxin (5). All cows progressed normally through the dry period to calving. Although direct infusion of bacteria into the mammary gland is one method of determining the protective value of a vaccine, it may not be the same as natural exposure. Entry of Gramnegative bacteria into the mammary gland occurs by passage of bacteria through the streak canal into the teat cistem (3). Direct deposition of E. coli into the cistern could bypass any protective mechanisms in the streak canal, eliminating the need for bacterial factors required to colonize or pass through the streak canal. Also, the number of bacteria gaining entrance into the teat or gland cistern may be very different than with naturally occurring
IMI. The current study demonstrated that the lactating mammary gland is not protected from experimental E. coli challenge following immunization with curli-producing strain of E. coli. Challenge with E. coli at 30 d of lactation resulted in clinical mastitis. All experimental IMI were self-eliiated and of short duration. ACKNOWLEDGMENTS
This study was supported in part by AlphaLaval AgriIntemational, Sweden. The authors thank Pamela Schoenberger, Sue Romig, Heidi Rennecker, Lucinda Shock, and Sandy Renner for technical assistance. Fibronectin binding data was generously supplied by Alfa-Laval AgriIntemational, Sweden. REFERENCES lcarroll, E. J., N. C. Jain, 0. W. Schalm, and J. Lasmanis. 1973. Experimentally induced coliform mastitis: inoculation of udders with serum-sensitive
825
and serum-resistant organisms. Am. J. Vet. Res. 3 4 1143. 2Crosby, W. H., J. 1. Munn, and F. W. FWh. 1954. Standardizing a method for clinical hemoglobinometry. US Amed Forces Med. J. 5:693. 3 Eberha~t,R. J., R. P. Natzke, F.H.S. Newbould, B. Nonnecke, and P. Thompson. 1979. Coliform mastitika review. J. Dairy Sci. 621. 4Erskine. R. I., R J. Eberhart, P. J. Grasso, and R W. Scholz. 1989. Induction of Escherichia coli mastitis in cows fed selcmum-deficient or selenium-supplemented diets. Am. J. Vet. Res. 502093. 5Frost, A. J., and B. E. Brooker, 1983. The effect of Eschm'chio coli endotoxin and culture filtrate on the dry bovine mammary gland. J. Comp. Pathol. 93:211. 6Frost, A. I., and A. W. Hill. 1982. Pathogenesis of experimental bovine mastitis following a small inoculum of Escherichia coli. Res. Vet. Sci. 33:105. 7Frost, A. I., D. D. Wauasi@e, and J. B. Woolcock 1977. Some factors affecting selective adherence of microorganisms m the bovine mammary gland. Infect. Immun. 15245. 8 Hill, A. W. 1981. Factors influencing the outcome of Escherichiu coli mastitis in the davy cow. Res. Vet. Sci. 31:107. 9 Hill,A. W., A. L. Shears,and K. G. Hibbitt. 1978. The elbination of serum-resistant Escherichiu coli from experimentally infected single mammary glands of healthy cows. Res. Vet. Sci. 25239. lOHogan, J. S., K. L. Smith,K. H. Hoblet, P. S. Schoenberger, D. A. Todhunter. W. D. Hueston, D. E. Pritchard, G. L. Bowman, L. E. Heider, B. L. Brockett, and H. R Conrad. 1989. Field swvey of clinical mastitis in low somatic cell count herds. J. Dairy Sci. 721547. 11 OMn A., A. Jonsson, and S. Normark. 1989. Fibronectin binding mediated by a novel class of surface organelles on Escherichio coli. Nature (Lond.) 33:652. 120lson, 0. E.,I. Palmer, and E. Cary. 1975. Modification of the official fluorometric method for selenium in plants. J. Assoc. Offic. Agric. Chem. 58:117. 13Opdebeeck, J. P.,A. J. Frost, and D. O'Boyle. 1988. Adhesion of Stuphylococcus uureus and Escherichiu coli to bovine udder epithelial cells. Vet. Microbiol. 16:77. 14Paglia, D. E., and W. N. Valentine. 1%7. Studies on the quantitative and qualitative characterizations of erythrocyte glutathione peroxidase. J. Lab. Clin. Med. 70158. 15 Rainard, P. 1983. Experimental mastitis with Escberichio coli: kinetics of bacteriostatic and bactericidal activities. Ann. Rech. Vet. 141. 16Sanchez-Carlo, V., J. S. McDonald, and R. A. Packer. 1984. Virulence factors of Escherichia coli isolated from cows with acute mastitis. Am. J. Vet. Res. 45: 1775.
17 Smith, K.L.,D. A. Todhunter, and P.S. Schoenberger. 1985. Environmental mastitis: cause, prevalence, prevention. J. Dairy Sci. 68:1531. 18Sokal, R. R., and F. J. Rohlf. 1969. Biometry. The principles and practice of statistics in biological research. W. H. Freeman and Company, San Prancisco, CA. Journal of Dairy Science Vol. 74, No. 3, 1991
