共查询到20条相似文献,搜索用时 31 毫秒
1.
B K?hler 《Archiv fuer experimentelle veterinaermedizin》1978,32(6):841-853
Nineteen Clostridium perfringens Type C strains and ten foreign control strains of subtypes C1, C3, and C4 were tested for their toxin formation and spore resistance to heat. The 19 Type C strains had been isolated from unweaned piglets in the context of necrotising enteritis outbreaks in the GDR. The Clostridium perfringens Type C strains formed beta-toxin, but they failed to form epsilon-toxin or gamma-toxin, alpha-toxin was successfully recorded from 15 of the 19 strains tested from unweaned piglets. The minor-lethal toxin fractions were also tested, with delta-toxin being recorded from all strains, non-alpha-delta-theta-toxin from six, theta-toxin from five, and K-toxin from one. Tests for delta-toxin, lambda-toxin, and mu-toxin were negative. The Clostridium perfringens Type C strains isolated in the GDR from unweaned piglets with necrotising enteritis were, basically, identical with those described in Denmark by von Hogh (1967) with regard to toxin formation. Clostridium perfringens strains cultured in broilers with necrotising enteritis were characterised by regular toxin production in the context of alpha, theta, delta as well as non-alpha-delta-theta. They failed to form beta, epsilon, gamma and lambda, while mu-toxin was formed by them quite irregularly. They, consequently, are Type A strains. Resistance to chloramphenicol and/or oxytetracycline was exhibited by 78.5 per cent of 237 tested Clostridium perfringens strains which had been isolated from unweaned piglets and broilers with necrotosing enteritis. Multiple resistance was recorded from 33.9 per cent. All strains were susceptible to penicillin, nitrofurantoin, and erythromycin. 相似文献
2.
B K?hler J Zabke R Sondermann H Pulst H J Rummler 《Archiv fuer experimentelle veterinaermedizin》1979,33(4):595-619
Necrotising enteritis had been the cause of death of 4.9 per cent in 5,177 nursed piglets, which was established by pathological examination. The number of piglets, in that context, which had come from industrialised sow breeding units was equivalent to 92 per cent. The nursed piglet held the third position, next to smaller ruminants (19.4 per cent) and fowl (6.0 per cent), with regard to the occurrence of Clostridium perfringens enterotoxemia or necrotising enteritis in 112,218 animals which were pathologically examined after death. Necrotising enteritis so far has been rare in the GDR. No regional accumulation has been observed. Several outbreaks on industrialised sow breeding units actually remained stationary. The occurrence of the disease may be favoured by a number of factors which are conducive to accumulation of Clostridium perfringens Type C in a given stock. Group keeping of pregnant sows, simultaneous farrowing of larger groups of sows, group treatment of nursed piglets, using neomycin, chloramphenicol, oxytetracycline, and other antibiotics to which Clostridium perfringens is primarily resistant or has acquired resistance in the course of time are some of those contributive factors. Transmission of Clostridium perfringens Type C through feedstuff is possible, though it would lead to a real outbreak only by high intensity of the contamination, and it played a minor role in proliferation of the disease. 3479 Clostridium perfringens strains were isolated from 9,481 animals, both clinically intact and after death, with 30 species being included. Type classification revealed 2454 strains of Type A (70 per cent), 204 of Type D (5.88 per cent), 164 of Type C (four per cent), and 48 of Type B (1.34 per cent). There were 688 atoxic strains (17 per cent). Swine is the major carrier of Clostridium perfringens Type C, with 87 per cent of all Clostridium perfringens Type C strains having been isolated from swine. Swine was followed by fowl (four per cent), sheep (four per cent), cattle, rabbit, and dog (1.27 per cent each). Clostridium perfringens Type C was obtained from the faeces of clinically intact sows in seven instances, including two cases with sows (0.46 per cent) from farms with no previous record of necrotising enteritis. 相似文献
3.
N Balauca B K?hler F Horsch R Jungmann E Prusas 《Archiv fuer experimentelle veterinaermedizin》1976,30(6):913-923
Coccidial oocysts and/or vegetative germs, spores, and toxins of Cl. perfringens Type A were used in mono-infection and poly-infection experiments on SPF chicken aged seven and 56 days and kept under different conditions. Necrotic enteritis was regularly reproduced in all experimental groups with polyinfections. In chicken aged seven days necrotic enteritis was reproduced even after repeated mono-infections with 4 X 10(9) gerus and toxin of Cl. perfringens. The loss figures recorded from ground-kept infected groups were higher than those established from the cage-kept animals. The pathologica-anatomic findings recorded from the dead chicken included necrotic, ulcerative, and catarrhal to haemorrhagic intestinal inflammations, with necrotic enteritis being most strongly pronounced in the ileum and jejunum. All infected groups lost not only animals but weight as well. Their average weight on the 21 st day from the onset of infection was up to 26.4 per cent lower than that of the control groups. 相似文献
4.
Prevalence of cpb2, encoding beta2 toxin,in Clostridium perfringens field isolates: correlation of genotype with phenotype 总被引:4,自引:0,他引:4
Beta2 toxin, encoded by the cpb2 gene, has been implicated in the pathogenesis of porcine, equine and bovine enteritis by type A Clostridium perfringens. By incorporating primers to cpb2 into a multiplex genotyping PCR, we screened 3270 field isolates of C. perfringens. Of these, 37.2% were PCR positive for the cpb2 gene. The majority of isolates from cases of porcine enteritis were positive for cpb2 (>85%), and this was even more true for C. perfringens isolated from cases of porcine neonatal enteritis (91.8%). In contrast, isolates from normal pigs only contained cpb2 in 11.1% of cases. The correlation between enteritis in other animal species and the presence of cpb2 was not so strong. cpb2 was found in 21.4% of C. perfringens isolates from cattle with enteritis, and in 47.3% of isolates from calves with enteritis or abomastitis. The prevalence of cpb2 varied with genotype, with type A isolates being positive for this gene in 35.1% of cases. Furthermore, enterotoxigenic type D or type E strains almost always carried cpb2. We cloned a 6xHIS-tagged beta2 (HIS-beta2) and used this protein to raise antiserum against beta2. Culture supernatants from 68 cpb2-positive and 13 cpb2-negative strains were tested for the presence of beta2 by Western blotting. In cpb2-positive isolates of porcine origin, beta2 was almost always detected (96.9%). However, in cpb2-positive isolates from other animal species, only 50.0% expressed beta2 protein. The high rate of cpb2-positivity among strains from neonatal pigs with enteritis and the high correlation of genotype with phenotype, supports the contention that beta2 toxin plays a role in the pathogenesis of these infections. However, it may be important to consider the use of an additional method for the detection of beta2 toxin in non-porcine cpb2-positive isolates when making claims about the role of beta2 in enteritis in non-porcine species. 相似文献
5.
Enterotoxigenic Clostridium perfringens type A isolated from intestinal contents of cattle, sheep and chickens. 总被引:2,自引:2,他引:0
下载免费PDF全文
![点击此处可从《Canadian journal of veterinary research》网站下载免费的PDF全文](/ch/ext_images/free.gif)
L Niilo 《Canadian journal of veterinary research》1978,42(3):357-363
One hundred and fourteen strains of Clostridium perfringens, isolated from the intestinal contents of cattle, sheep, and chickens with enteritis or other disease conditions were studied for their ability to produce enterotoxin. Reversed passive hemagglutination, fluorescent antibody and immunodiffusion tests were used. On the basis of the reversed passive hemagglutination titres, supported by the other two tests, enterotoxigenicity of the strains was arbitrarily classified into two categories: highly enterotoxigenic and potentially enterotoxigenic, with 12% falling into each category. All the highly enterotoxigenic strains originated from cases of enteritis and included all three animal species. Apart from enterotoxigenicity, one C. perfringens strain produced beta toxin (type C) and 21 strains produced large amounts of alpha-toxin. The latter strains were predominantly associated with necrotic enteritis in chickens. 相似文献
6.
Gut P Luginbühl A Nicolet J Boerlin P Burnens AP 《Schweizer Archiv für Tierheilkunde》2002,144(6):275-281
Investigations were performed on shedding of C. perfringens in sows from four different pig farms. In two farms where no outbreaks of necrotizing enteritis had been observed, no strains of C. perfringens producing beta-toxin were detected in the faeces of sows. In contrast, C. perfringens strains producing beta-toxin were detected in sows on both farms suffering outbreaks of acute necrotizing enteritis. Strains of C. perfringens producing beta-toxin were invariably positive for the beta 2-toxin gene. However, strains carrying the beta 2-toxin gene only (i.e. negative for beta-toxin) were present in animals on all farms with roughly similar frequencies (mean 28.2% carriers). Some sows carried C. perfringens strains of both toxin genotypes simultaneously. Whereas these data further support the role of betatoxin as a cause of necrotizing enteritis, the role of beta 2-toxin in intestinal disease of piglets remains unclear. To establish the role of faecal shedding vs. environmental contamination as reservoirs of C. perfringens type C, strains were isolated from teats and feedlot trough swabs (toxin genotype beta/beta 2), as well as from fodder (genotype beta 2). However, sows carried this pathogen intermittently and in small numbers. This renders an individual, reliable diagnosis of carrier sows very difficult. Ribotyping of 34 C. perfringens isolates of different toxin genotypes showed five distinct profiles. Different toxin genotypes can belong to the same ribotype, and the same toxin genotype can be present in different ribotypes. Thus, even if a majority (79.4%) of strains investigated in a limited geographic region belonged to ribotype 1, ribotyping offered discrimination of strains beyond toxin typing. 相似文献
7.
Clostridial enteric infections in pigs. 总被引:1,自引:0,他引:1
Clostridium perfringens types A and C and Clostridium difficile are the principal enteric clostridial pathogens of swine. History, clinical signs of disease, and gross and microscopic findings form the basis for a presumptive diagnosis of C. perfringens type-C enteritis. Confirmation is based on isolation of large numbers of type-C C. perfringens and/or detection of beta toxin in intestinal contents. Diagnosis of C. perfringens type-A infection, however, remains controversial, mostly because the condition has not been well defined and because type-A organisms and their most important major (alpha) toxin can be found in intestinal contents of healthy and diseased pigs. Isolation of large numbers of C. perfringens type A from intestinal contents, in the absence of other enteric pathogens, is the most reliable criterion on which to base a diagnosis. Recently, beta2 (CPB2) toxin-producing C. perfringens type A has been linked to disease in piglets and other animals. However, implication of CPB2 in pathogenesis of porcine infections is based principally on isolation of C. perfringens carrying cpb2, the gene encoding CPB2, and the specific role of CPB2 in enteric disease of pigs remains to be fully defined. Clostridium difficile can also be a normal inhabitant of the intestine of healthy pigs, and diagnosis of enteric infection with this microorganism is based on detection of its toxins in feces or intestinal contents. 相似文献
8.
S M Shane J E Gyimah K S Harrington T G Snider 《Veterinary research communications》1985,9(4):269-287
Sporulated oocysts of Eimeria acervulina were administered orally to cage-housed broilers at a dose of 3.5 X 10(5) resulted in mild subclinical coccidiosis. Clostridium perfringens incorporated in feed at a level of 2.5 X 10(8) organisms/g. produced lesions characteristic of necrotic enteritis. Mortality of 8% (7/80) occurred in birds fed a ration inoculated with Cl. perfringens alone. Mortality of 35% (28/80) was observed in birds which received an oral dose of E. acervulina and which were fed simultaneously with a ration containing Cl. perfringens. Birds which were fed an inoculated ration two days after an oral dose of E. acervulina showed 41% (33/80) mortality. Birds which received an inoculated ration for two days before administration of an oral dose of E. acervulina demonstrated 18% mortality (15/80). Birds which were fed an inoculated ration four days after an oral dose of E. acervulina showed 10% mortality. Infection with E. acervulina reduced the pH of intestinal contents with a simultaneous depression in serum protein. A 39% increase in intestinal passage time from 178 to 248 minutes occurred on the fifth day after infection with E. acervulina. These experiments suggest that necrotic enteritis, attributed to proliferation of a toxigenic strain of Cl. perfringens, followed intestinal stasis and minimal lesions induced by mild intestinal coccidiosis. 相似文献
9.
B K?hler J Zabke K Reiher H J Rummler 《Archiv fuer experimentelle veterinaermedizin》1979,33(4):621-637
Recent methods used and experience obtained in the control of necrotising enteritis are reported in this paper, with reference being made to both the pathogenesis and epizootiology of the disease. Two inoculations of the sows, using "Enterotox?mievakzine Dessau bivalent" five and three weeks before parturition, have worked well for prophylaxis. Oral treatment was applied to nursed piglets, using 40,000 I.U. of "Aviapen" and "V-Tablopen" penicillin per animal and day over periods between two and four days, helped to minimise piglet loss, particularly in the period between a fresh outbreak and full effectiveness of immunoprophylactic action. Such treatment was conducted metaphylactically and therapeutically. The first metaphylactic treatment was given within 24 hours from parturition. Combination of mother animal vaccination with the above therapeutic use of those two penicillin preparations worked extremely well in enzootically contaminated stocks and proved to be the most effective approach, for the time being, to controlling necrotising enteritis of nursed piglets. Yet, all those control measures failed to bring about full stock sanitation on industrialised units. Sow trading was not permitted until at least four weeks had elapsed from full effectiveness of mother animal vaccination, with the view to reducing the proliferation of Clostridium perfringens Type C via sales of breeding animals. All sows were given two "Enterotox?mievakzine Dessau bivalent" vaccinations, prior to sale. The animals were sold only to smaller farms (less than 500 sows for breeding) with concentional keeping patterns which were kept under constant diagnostic supervision. Neomycin, oxytetracycline, chloramphenicol, and other antibiotics against which Clostridium perfringens was resistant or in a position to assume resistance were used on endangered stocks only in conjunction with penicillin or not at all. This programme of control has proved to be efficient through a period of more than three years. 相似文献
10.
11.
Toxinotypes of Clostridium perfringens isolated from sick and healthy avian species. 总被引:1,自引:0,他引:1
Rocio Crespo Derek J Fisher H L Shivaprasad Mariano E Fernández-Miyakawa Francisco A Uzal 《Journal of veterinary diagnostic investigation》2007,19(3):329-333
Currently, the factors/toxins responsible for Clostridium perfringens-associated avian enteritis are not well understood. To assess whether specific C. perfringens' toxinotypes are associated with avian enteritis, the isolates of C. perfringens from 31 cases of avian necrotic or ulcerative enteritis submitted between 1997 and 2005 were selected for retrospective analysis using multiplex PCR. C. perfringens was isolated from chickens, turkeys, quail, and psittacines. The toxinotypes of isolates from diseased birds were compared against the toxinotype of 19 C. perfringens isolates from avian cases with no evidence of clostridial enteritis. All C. perfringens isolates were classified as type A regardless of species or disease history. Although many isolates (from all avian groups) had the gene encoding the C. perfirngens beta2 toxin, only 54% produced the toxin in vitro when measured using Western blot analysis. Surprisingly, a large number of healthy birds (90%) carried CPB2-producing isolates, whereas over half of the cpb2-positive isolates from diseased birds failed to produce CPB2. These data from this investigation do not suggest a causal relationship between beta2 toxin and necrotic enteritis in birds. 相似文献
12.
Toxigenic characteristics of Clostridium perfringens type C in enterotoxemia of domestic animals. 总被引:2,自引:0,他引:2
下载免费PDF全文
![点击此处可从《Canadian journal of veterinary research》网站下载免费的PDF全文](/ch/ext_images/free.gif)
L Niilo 《Canadian journal of veterinary research》1987,51(2):224-228
Eleven Clostridium perfringens type C strains isolated from fatal cases of hemorrhagic enterotoxemia of Canadian calves, a piglet, and a foal were studied for the production of soluble antigens. All the isolates from calves and a foal failed to produce delta toxin, but were capable of producing large amounts of lethal beta toxin. A strain isolated from a piglet produced delta, but very little beta toxin. Other differences were relatively minor. The results indicated that young domestic animals may be susceptible to all subtypes of C. perfringens type C. A simple method of using blood agar plates coated with type A antiserum for demonstration of hemolytic patterns was found advantageous in differentiation of C. perfringens strains. 相似文献
13.
Porcine Clostridium perfringens type A spores, enterotoxin and antibody to enterotoxin 总被引:2,自引:0,他引:2
Forty-two Clostridium perfringens type A strains isolated from cases of diarrhoea in pigs were tested for their ability to sporulate and produce enterotoxin in three different sporulation media. Enterotoxin was produced by 11 of the 42 C perfringens type A isolates (26.2 per cent). Thirteen isolates (30.9 per cent) produced spores at a frequency of 10 per cent or more. Spore production was recorded in 24 (57.1 per cent) of the isolates. The titres of enterotoxin produced by the isolates ranged from 1:2 to 1:64. The enterotoxin produced was compared with that produced by a reference strain and found to be identical. Ninety-eight of 106 sow sera from four different farms were found to possess antibodies to C perfringens type A enterotoxin with titres ranging from 1:2 to 1:64. Spores of C perfringens type A were detected in pig faeces and intestinal contents in 20 of 23 cases of enteritis at levels of up to 5 x 10(6) cells/g of faeces. Smaller numbers of spores, up to 2 x 10(4)/g were present in five of 10 samples from non-diarrhoeic pigs. Enterotoxin was demonstrated by Vero cell assay in five of the 23 samples from diarrhoeic pigs but in none of the 10 samples from non-diarrhoeic animals. It was clear from these studies that C perfringens type A strains in pigs could sporulate and produce enterotoxin in vitro and in vivo and that enteritis might be associated with sporulating organisms in vivo. 相似文献
14.
Garcia JP Beingesser J Fisher DJ Sayeed S McClane BA Posthaus H Uzal FA 《Veterinary microbiology》2012,157(3-4):412-419
Clostridium perfringens type C is an important cause of enteritis and/or enterocolitis in several animal species, including pigs, sheep, goats, horses and humans. The disease is a classic enterotoxemia and the enteric lesions and associated systemic effects are thought to be caused primarily by beta toxin (CPB), one of two typing toxins produced by C. perfringens type C. This has been demonstrated recently by fulfilling molecular Koch's postulates in rabbits and mice. We present here an experimental study to fulfill these postulates in goats, a natural host of C. perfringens type C disease. Nine healthy male or female Anglo Nubian goat kids were inoculated with the virulent C. perfringens type C wild-type strain CN3685, an isogenic CPB null mutant or a strain where the cpb null mutation had been reversed. Three goats inoculated with the wild-type strain presented abdominal pain, hemorrhagic diarrhea, necrotizing enterocolitis, pulmonary edema, hydropericardium and death within 24h of inoculation. Two goats inoculated with the CPB null mutant and two goats inoculated with sterile culture media (negative controls) remained clinically healthy during 24h after inoculation and no gross or histological abnormalities were observed in the tissues of any of them. Reversal of the null mutation to partially restore CPB production also increased virulence; 2 goats inoculated with this reversed mutant presented clinical and pathological changes similar to those observed in goats inoculated with the wild-type strain, except that spontaneous death was not observed. These results indicate that CPB is required for C. perfringens type C to induce disease in goats, supporting a key role for this toxin in natural C. perfringens type C disease pathogenesis. 相似文献
15.
16.
Association of genes encoding beta2 toxin and a collagen binding protein in Clostridium perfringens isolates of porcine origin 总被引:1,自引:0,他引:1
Clostridium perfringens is a cause of economically significant enteritis in livestock. Beta2 toxin, encoded by one of two cpb2 alleles, is implicated as a virulence factor in this disease. Previous studies determined that the consensus cpb2 allele is preferentially associated with C. perfringens isolated from pigs. In C. perfringens strain 13, the consensus cpb2 allele is found on the plasmid pCP13, which also carries cna, encoding a putative collagen binding protein, CpCna. This protein was shown to be a bona fide collagen adhesin, as recombinant, HIS-tagged CpCna bound collagen type I as determined by Far Western blotting. Genomic DNA from C. perfringens isolated from a variety of host species were subjected to PCR to determine the prevalence of cna in these strains and correlate its carriage with the presence and type of cpb2 allele. The cna gene was found in 55.8% of isolates from all host species (n=208) and 68.1% of porcine isolates (n=119). In cpb2+ isolates, cna was present in 69.9% of isolates from all hosts (n=153), but was found in 98.7% of porcine isolates (n=75). Furthermore in porcine isolates, the consensus cpb2 allele and cna were absolutely correlated with the presence of pcp12, a pCP13-encoded gene, and pcp12 was never found in any isolate that lacks either cpb2 allele. The finding that CpCna binds collagen and that the cna gene is associated with the consensus cpb2 allele implicates CpCna as a potential virulence factor in porcine enteritis caused by C. perfringens. 相似文献
17.
Cytotoxins in non-enterotoxigenic strains of Escherichia coli isolated from feces of diarrheic calves 总被引:10,自引:0,他引:10
We have examined the cytotoxic responses produced in HeLa and Vero cell cultures by sonicates from 15 non-enterotoxigenic (STa-, LT-) strains of E. coli, highly lethal for mice parenterally LD50 less than 3 X 10(7) CFU), which had been isolated from feces of diarrheic calves. Three types of cytotoxic responses were observed. Type 1 (five strains) consisted of enlargement, rounding and polynucleation of HeLa cells, an effect previously reported with cytotoxic necrotizing factor (CNF) in E. coli from infant and piglet enteritis. Type 2 toxicity (three strains and the control Vir strain S5) was also characterized by enlargement and polynucleation of HeLa cells, but in contrast to Type 2 effect, cells were elongated. Sonicates from the latter strains were lethal for chickens, producing the lesions previously described with Vir strains. Type 3 toxicity (two strains and the control VT strain H19), produced an extensive destruction of both Vero and HeLa cell cultures. Cytotoxic effects were completely abolished upon heating for 1 h at 60 degrees C for Type 1 and 2 extracts and at 80 degrees C for Type 3 extracts. Seroneutralization assays showed that cytotoxins of the same type were closely related antigenically. In addition, a slight cross-neutralization was observed between Type 1 (CNF) and Type 2 (Vir) toxins. 相似文献
18.
19.
Cai Y Gao J Wang X Chai T Zhang X Duan H Jiang S Zucker BA Schlenker G 《DTW. Deutsche tier?rztliche Wochenschrift》2008,115(8):292-4, 296-7
Four hundred and twenty intestinal content samples (not including intestinal tissues) of freshwater fishes (60 silver carps, 100 carps, 100 crucian carps, 60 catfishes and 100 zaieuws) caught from one water reservoir were examined bacteriologically for the occurrence of C. perfringens. Isolates were examined by polymerase chain reaction (PCR) for genes encoding the four lethal toxins (alpha, beta, epsilon and iota) for classification into toxin types and for genes encoding enterotoxin and the novel beta2 toxin for further subclassification. C. perfringens could be isolated in 75 intestinal contents samples (17.9%) from freshwater fish including: 13 silver carps, 2 carps, 12 crucian carps, 40 zaieuws, and 8 catfishes. In 75 isolates, 58 strains (77.3%) were C. perfringens toxin type C (alpha and beta toxin positive), 13 strains (17.3%) were toxin type A (alpha toxin positive) and 4 strains (5.3%) were toxin type B (alpha, beta and epsilon toxin positive). In addition, the gene encoding for beta2 toxin was found in 47 strains (62.7%) of all the isolates, seven from type A, two from type B, and 38 from type C. The gene encoding for enterotoxin was not found in any isolate. These amplified toxin gene fragment were cloned and sequenced and compared with reference strains, the identity varied from 98.15% to 99.29%. This is the first report of C. perfringens alpha, beta, epsilon, beta2 toxins in freshwater fish and of beta, epsilon toxins in fish in general, and is the first discovery that the beta2 toxin could be detected in strains of type B. The origin of this bacterium and its importance to human food poisoning in freshwater fish is discussed. 相似文献
20.
Timbermont L Lanckriet A Gholamiandehkordi AR Pasmans F Martel A Haesebrouck F Ducatelle R Van Immerseel F 《Comparative immunology, microbiology and infectious diseases》2009,32(6):503-512
Since the ban on growth-promoting antibiotics in animal feed in the European Union, necrotic enteritis has become a major cause of mortality in broiler chickens. Despite the importance of the disease, the pathogenesis is still not completely understood. In the current study, Clostridium perfringens strains isolated from healthy flocks and isolates from outbreaks of necrotic enteritis were evaluated for the ability to cause gut necrosis in an intestinal loop model in laying hens and in an experimental infection model in broilers. High, intermediate and low alpha toxin producing strains were chosen from each isolation source. Only the isolates from field outbreaks induced necrotic gut lesions, independent of the amount of alpha toxin produced in vitro. It was also shown that alpha toxin producing isolates from calf hemorrhagic enteritis cases were not able to induce necrotic enteritis in poultry. These results suggest the presence of host specific virulence factors in C. perfringens strains, isolated from chickens with intestinal necrotic enteritis lesions. 相似文献