Gut, 1988, 29,1222-1228

Investigation of mycobacteria in Crohn's disease tissueby Southern blotting and DNA hybridisation withcloned mycobacterial genomic DNA probes from aCrohn's disease isolated mycobacteriaP D BUTCHER, J J McFADDEN, AND J HERMON-TAYLOR

From the Department ofSurgery, St George's Hospital Medical School, London

SUMMARY A mycobacterial aetiology for Crohn's disease (CD) has been suggested. Slow growingmycobacteria indistinguishable from M paratuberculosis, the causative agent of enteritis inruminants (Johne's disease) have been isolated from CD tissues. We have used cloned genomic DNAprobes derived from a CD isolated mycobacteria strain Ben, to investigate the presence ofmycobacterial DNA sequences in CD tissues. DNA was extracted from total tissue from 17 CD andfour control gut specimens. DNA was digested with restriction endonucleases, electrophoresed andtransferred to nylon membranes by Southern blotting and hybridised to radiolabelled DNA probes.No mycobacterial DNA was detected in any tissue sample studied. Reconstitution experiments withknown numbers of in vitro cultured mycobacteria showed sensitive detection of mycobacterial DNA.DNA extracted from mouse liver, infected with M lepraemurium revealed a strong hybridisationsignal and showed the applicability of the experimental approach to the detection of mycobacterialDNA in naturally infected tissues. The results do not provide evidence for the involvement ofmycobacteria in the pathogenesis ofCD but do not exclude the possibility of low levels of infection insubsets of intestinal cells with spheroplast or cell wall deficient forms of mycobacteria.

An infectious aetiology for Crohn's disease (CD) hasoften been suggested. '- Macroscopic and histologicalsimilarities between CD, intestinal tuberculosis inman and paratuberculosis in ruminants (Johne'sdisease) caused by Mycobacterium paratuberculosis67have led to the suggestion that mycobactena may alsobe involved in CD,'9 even though earlier studiesfailed to show the presence of acid fast myco-bacteria.'" Burnham et al'2 have isolated, afterprolonged culture, a mycobacterium similar to Mkansasii from a lymph node of one out of 27 cases ofCD, as well as unidentified pleomorphic, cell walldeficient organisms (spheroplasts) from 22 of theseCD cases, and one of 11 non-inflammatory diseasecontrols. They suggested that mycobacterial sphero-plasts might be an aetiological agent in CD. Sphero-

Address for correspondence: Dr P D Butcher, Department of Gastro-enterology, St Bartholomew's Hospital, West Smithfield, London ECIA 7BE.

Received for publication 15 April 1988.

plast forms of common bacteria have been found invarious inflammatory diseases of the intestine.'3A slow growing, mycobactin (an iron chelating

growth factor) dependent Mycobacterium specieshas recently been isolated from three of 14 CD guttissues, but not controls, after 18 months culture.'4This organism was similar to M paratuberculosis andoral inoculation of neonatal goats with these isolatesproduced granulomatous, non-caseating tuberculoidileitis within three to five months'5 from whichmycobacteria were reisolated. 16 Subsequently, afurther similar isolate has been recovered from CDtissue. '` These CD mycobacterial isolates wereinitially recovered in primary culture as slow growingspheroplasts which reverted to bacillary form onlyafter prolonged and repeated subculture. Thespheroplast ancestry of the CD bacillary isolates wasconfirmed by the analysis of restriction fragmentlength polymorphisms of ribosomal genes by DNAhybridisation. '1 We have also shown that these CD

1222

on Decem

ber 21, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.29.9.1222 on 1 Septem

ber 1988. Dow

nloaded from

Investigation ofmycobacteria in Crohn's disease tissue

isolated mycobacteria are identical to each other andto M paratuberculosis by the use of DNA probeswhich identify restriction fragment length poly-morphisms. "20 Unidentified spheroplasts wererecovered from 61% of CD tissues but not controls.'7Several other groups have recently reported indepen-dently the isolation of mycobacteria from CD tissuesof types similar to M paratuberculosis as well as toM chelonei and M avium.22The repeated isolation from CD tissues of a

mycobacterial pathogen may be of aetiologicalsignificance in CD. The fastidious growth require-ments, however, and long term cultures of CDderived organisms, makes their study protracted andvery difficult. Their possible presence as spheroplastsor cell wall deficient forms also precludes theirdetection in CD tissue by microscopy. Direct demon-stration of mycobacterial forms in CD tissues musttherefore rely on alternative techniques. Detectionof mycobacterial genomes by DNA hybridisationwould provide a rapid and sensitive approach to thisproblem. A recent report of DNA-DNA hybridisa-tion in solution of CD tissue total DNA with totalDNA from a CD mycobacterial isolate demonstratedmycobacterial DNA sequences in both CD andcontrol gut tissue.24

Precise identification of mycobacteria in CD tissueis necessary, however, for their aetiological signific-ance to be evaluated, particularly if closely relatednon-pathogenic species such as M avium may befound in control tissue. We have therefore used amore specific method to detect DNA sequences ofmycobacterial origin that makes use of clonedgenomic DNA probes derived from a CD myco-bacterial isolate, strain Ben identified as M para-tuberculosis.'4 1 These probes are Mycobacteriumspecific and have been used to differentiate betweenmycobacterial species by restriction fragment lengthpolymorphism analysis.202' We report here the use ofcloned DNA probes to attempt to detect myco-bacterial DNA in tissue derived DNA from 17 CDand four control gut tissues by Southern blotting afterrestriction endonuclease digestion of test DNA andhybridisation with radiolabelled probes.

Methods

TISSUE SOURCE

Specimens of colon and terminal ileum wereremoved under sterile conditions from Crohn'sdisease patients undergoing surgical resection at StGeorge's Hospital between September 1983 andJune 1986. Full thickness intestinal specimens weretaken from involved areas showing typical thicken-ing, cobblestoning and ulceration. All cases of CDwere confirmed by clinical, radiological, and histo-

logical criteria. Disease control tissues were fromcases of colon carcinoma. Specimens were trimmedof fat, washed in sterile saline and snap frozen inliquid nitrogen and stored at -70°C. The totalnumber of intestinal specimens screened for all formsof mycobacteria in this report were 17 CD (twocolon, one jejunum, 14 terminal ileum) and fourcontrols (one colon, three terminal ileum).

DNA EXTRACTIONDNA was extracted from tissue homogenates by anenzymic procedure2` shown to extract DNA frombacillary mycobacteria which are resistant to lysis byconventional DNA extraction procedures. Briefly,tissue in TEN buffer (50mM Tris-HCI, pH8, 150mMNaCl, 100 mM EDTA) was sequentially digestedwith 10 mg/ml Subtilisin (Sigma; protease Type VIII)at 37°C for three hours, 50 mg/ml lysozyme at 50°Cfor three hours and 3 mg/ml pronase (Calbiochem) at37°C for 12 hours after the addition of 1% SDS. DNArecovered by phenollchloroform extraction andethanol precipitation was redissolved in TE (10 mMTris-HCI, pH 8, 1 mM EDTA) and treated with 10Wg/ml RNA'ase A for two hours at 37°C after heating

to 65°C for 15 minutes. DNA was again phenol/chloroform extracted, ethanol precipitated from 0* 1M NaCI and redissolved in TE buffer. DNA concen-tration was measured by absorbance at 260 nm(1 A260 unit=50 [sg/ml) and its integrity checked byagarose gel electrophoresis. High molecular weightDNA was obtained in all cases.

PROBESCloned mycobacteria specific probes used in thisstudy were from the CD derived M paratuberculosisisolate, strain Ben'4. BamHl fragments of genomicDNA were cloned into the plasmid vector pGEM-1as previously described.20 Probes were selected fortheir ability to detect and discriminate between arange of mycobacteria on the basis of restrictionfragment length polymorphisms (RFLP). Threeclones were used: (i) pMB22; with a 5 kb insertincluding a 1-2 kb repetitive element within it that isrepeated greater than 10 times inMparatuberculosis;(ii) pMB21; this clone hybridises equally well to Mparatuberculosis, M avium, CD isolates, and less toM kansasii; (iii) pMB 19; this clone hybridises equallywell to M paratuberculosis, CD isolates, M avium,and M kansasii The clone pMB19 has also beenshown to hybridise to M bovis, M tuberculosis, Mintracellulare, and M chelonei. These clones and theRFLP's they generate have been described in detailpreviously.2`21

Radiolabelled DNA probes were prepared forhybridisation by BamHl excision of the cloned insertand separation by electrophoresis in low temperature

1223

on Decem

ber 21, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.29.9.1222 on 1 Septem

ber 1988. Dow

nloaded from

Butcher, McFadden, and Hermon- Taylor

8 91011121314 1 2345 6 7 8 91011121314

Fig. 1 Specificity and sensitivity ofdetection ofmycobacterial DNA sequences using cloned genomic DNA probes. Lanes1-5 (respectively): total DNA, 1jig, from normal human gut tissue reconstituted with known numbers ofM avium bacilli:5x 106, 2x 10i, 2x 106, 5x104, I x10'. Lanes 6 and 7: 10 [tg DNA from mouse liver infected with M lepraemurium. Lane 8:

10 p.g normal human leucocyte DNA. Lanes 9-12 (respectively): 1, 5, 20, and 50 ng DNA from M avium bacilli. Lanes 13 and14: 10 ng ofDNA from M paratuberculosis and 200 ng ofDNA from M phlci, respectively. DNA was electrophoresed afterPVU-1 1 restriction endonuclease digestion, Southern blotted and hybridised consecutively with thefollowing probes: (a)pMB22 and (b) pMB2J and autoradiographed. Molecular weight size markers are indicated (-) and arefrom the top ofthegel 21, 9.4, 6-6, 4 4, 2-3, and 2-0 kilobase pairs (kb).

gelling agarose; 20-25 ng were radiolabelled with 50UCi [32p] deoxy CTP (Amersham; 3000 Ci/mmol) byhexanucleotide random priming and Klenow exten-sion.26 Double stranded DNA probe was denaturedby boiling before hybridisation. Specific activities of4-1-9-8x 106 cpm/tg DNA were obtained withapproximately 80% incorporation of the label.

HYBRIDISATION

Extracted DNA was digested with restriction endo-nuclease Hinfl and PVU-11 and 10 [tg electro-phoresed on 1% agarose gels, denatured in 0(4 MNaOH, neutralised and transferred by Southernblotting to nylon filters (Amersham; Hybond-N)following standard protocols.27 DNA was cross-linked to the filters by ultraviolet irradiation and thefilters prehybridised for 18 hours at 65°C in hybridisa-

tion solution: 3xSSC (1 xSSC=0-15 M NaCl, 0-015M Na citrate, pH 7.0), 10 mM NaH2PO4, pH 7, 05%SDS, 0-05% BSA, 0-05% polyvinyl pyrrolidine,0-05% Ficol, 10% dextran sulphate and 110 [sg/mldenatured salmon sperm DNA. Hybridisation wascontinued by the addition of the boiled radiolabelledprobe for 18 hours at 65°C. Filters were then removedand washed at 65°C in 3xSSC, 0-1% SDS for 3x15minutes and then 2x30 minutes in lxSSC, 0)l%SDS. Hybridisation of the labelled probe with targetDNA was detected by autoradiography at -70°Cwith intensifying screens for up to 30 days. Filterswere reprobed with a fresh probe after removal ofprevious probe with 0-4 M NaOH at 45°C for 45minutes followed by neutralisation in 0(2 M Tris-HCI, pH 7-4, lxSSC. In this way filters can bereprobed repeatedly without loss of signal.

1 2345 6 7

1224

on Decem

ber 21, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.29.9.1222 on 1 Septem

ber 1988. Dow

nloaded from

Investigation ofmycobacteria in Crohn 's disease tissue

Results

SENSITIVITY AND APPLICABILITY OF THEEXPERIMENTAL APPROACHIn order to quantify the sensitivity of this approach tthe detection of mycobacteria in CD tissue, a recoistitution experiment was done in which dilutions (

freshly cultured M avium ranging from 109 to 11bacilli were added to 0-1 ml of total control gut tissuhomogenate (equivalent to 0-03 g of tissue). M aviuiDNA was detected after coextraction with tissuDNA by hybridisation with the DNA probeApproximately 50 ,ug DNA was recovered from eacextraction, and 1 [tg was loaded in each lane felectrophoresis. The results are shown in FigureThe minimal detectable signal was obtained wit2x105 bacilli (lane 3, Fig. la and b). A separalexperiment in which known amounts of pure Xavium DNA, digested with PVU-11, were electriphoresed and hybridised with probes pMB21 an

1 2 3 4 5 6 7 8 9 10

4;. .. ...

Table 1 Limits of detection of Mycobacterial DNA usingthe three cloned genomic DNA probes: expressed as onemycobacterial genome per number ofhuman cells

Probe

Mycobacterial species pMB22 pMB21 pMBI9

M paratuberculosis 1/100 1/10 1/10M avium-intracellulare 1/10 1/10 1/10M kansasii 1/5 1/5 1/10

pMB22 showed that the minimal detectable amountof DNA was 1 ng after autoradiography; see (Fig. laand b lanes 9-12). The signal obtained from 1 ngmycobacterial DNA was routinely visible against thenegligible hybridisation background of 10 ,ug humanDNA as may be observed in Fig. lb, lane 8 and Fig. 3.Based on genome sizes of 3-1 x109 daltons for Mavium2' and 3x 1012 daltons for humans, this was

1 2 3 4 5 6 7 8 9 10

Fig. 2 Screening ofCrohn 's disease and control gut tissue for mycobacterial DNA. 10 p.g total DNA from terminal ileum ofeight cases of Crohn's disease (lanes 1-8) and I control (lane 9), digested with PVU-11 (a) or with Hinfl (b) restrictionendonuclease and hybridised with probepMB22. Lane 10: 10 ng M paratuberculosis DNA. Molecular weight markers (- ) asin Fig. 1.

1225

on Decem

ber 21, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.29.9.1222 on 1 Septem

ber 1988. Dow

nloaded from

Butcher, McFadden, and Hermon- Taylor

1 2 3456789101112 13141516

Fig. 3 Screening of Crohn's disease and control gut tissuefor mycobacterial DNA. 10 [tg total DNA extracted from theterminal ileum ofa further nine samples of Crohn 's diseasegut tissue (lanes 1-9) and three cases ofcolon carcinoma(lanes 10-12), digested with PVU 11 and hybridised with theprobepMB21. Lane 13: blank. Lanes 14 & 16: 1 ng M aviumDNA. Lane 15: 1 ng M paratuberculosis DNA. Molecularweight markers (-) as Fig. 1.

calculated to represent 2x 105 bacilli. Against a

background of 10 ,ug human DNA, this represents a

detection limit of 1 mycobacterial genome/10 humancells.The strength of hybridisation signal for M para-

tuberculosis DNA with clone pMB22 compared withclone pMB21 (or pMB19 - not shown) shows a 10fold increase, Fig. la and b, lanes 13. This is becausethe target sequence for pMB22 is an insertionalelement repeated greater than 10 times in thegenome of M paratuberculosis." The detection limitusing pMB22, for mycobacteria similar to M para-

tuberculosis already isolated from CD tissues istherefore one Mycobacterium/100 human cells. Thedetection limit for different mycobacteria speciesusing the three probes, presented in the Table, arebased on these and previous studies.'" 21 Figure 1 alsoshows the RFLP patterns obtained with differentspecies and shows that M avium, M paratuberculosis,and M phlei are clearly and precisely distinguishable.To show the applicability of this experimental

approach to detecting naturally occurring myco-

bacterial infections we applied these techniquesto 40 mg liver from a mouse infected with M

lepraemurium. Hybridisation analysis with clonespMB22 and pMB21 showed a very strong signal afterautoradiography, Fig. la and b, lanes 6 and 7. Thesignal obtained was at least as strong as that from5x 10' M avium bacilli, which against a backgroundof 10 .tg mouse DNA represents an infection levelof approximately 2.5 organisms/mouse liver cell.Clearly from Figure 1, lanes 1-3, at least a 25-foldreduction in signal would still be detectable, indicat-ing a level of sensitivity of detection for Mlepraemurium of one organism/10 cells.

DETECTION OF MYCOBACTERIAL DNA WITHIN CD

INTESTINAL DNA EXTRACTSDNA was extracted from full thickness involvedintestinal specimens from a total of 17 cases of CDand four normal control colon carcinoma cases. Amaximum of 10 [ig restriction endonuclease digestedDNA (equivalent to DNA from approximately 6 mgof tissue) was electrophoresed in each lane, blottedonto nylon filters and hybridised consecutively withthe three cloned probes, as described. As internalcontrols for hybridisation sensitivity, 10 ng ofrestricted DNA from M paratuberculosis was also runon each gel. No bands were observed for any of thetissue samples studied with any of the probes. Thecontrol lanes gave a strong signal with each probe.Typical results are illustrated in Figure 2 which showshybridisation of pMB22 with Hinfl and PVU-11digestions of DNA from eight CD and one controlspecimens. This experiment was repeated with afurther nine CD and three control tissues probed withall three probes after digestion with PVU-11 andHinfl and gave similar negative results. Figure 3shows the absence of mycobacterial sequences inthese tissues using probe pMB21 after PVU-1 1digestion by comparison with 1 ng DNA from Mavium and M paratuberculosis (lanes 14-16). Inconclusion, M paratuberculosis DNA could not bedetected at levels equivalent to one mycobacterial

I genome/100 human cells nor M avium and M kansasiiat one genome/10 human cells.

Discussion

Using cloned genomic DNA probes generated from aCD-derived strain of M paratuberculosis, strainBen'4, we were unable to detect mycobacterial DNAsequences in the DNA extracted from CD gut tissues.

t The probes were shown to be capable of sensitivedetection of mycobacterial DNA and of discriminat-ing between closely related mycobacteria by restric-

l tion fragment length polymorphism analysis`'(RFLP). The combination of the three probes used inthis study would have detected DNA related to Mparatuberculosis, M av,ium-intracellulare, and M

1226

on Decem

ber 21, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.29.9.1222 on 1 Septem

ber 1988. Dow

nloaded from

Investigation ofmycobacteria in Crohn's disease tissue 1227

kansasii, all species that have been variously isolatedfrom CD tissues. The maximum levels of sensitivityachieved for the detection ofM paratuberculosis wasone organism/100 human cells and for the otherspecies one organism/10 cells. Whether such levelsmight be similar to natural infections by theseorganisms is unknown. Certainly acid fast bacilli atthese levels have not been detected by staining.'°0Cell wall deficient or spheroplast forms of myco-bacteria which have been recovered from cultures ofCrohn's disease tissue by several groups"142223 maybe the natural intracellular form of mycobacteria inCD, as previously suggested.'2 There are, however,no data available for the intracellular abundance ofthese forms of mycobacteria. We suggest that if theseforms do occur in CD then they persist at levelsbeyond our present methods of detection in subsetsof cells permissive to invasion by and persistence ofmycobacteria, such as macrophages or giant cellswithin granulomata.

Although extraction methods and sensitivity calcu-lations were based on in vitro cultured M aviumbacilli, M lepraemurium infected mouse liver pro-vided a positive control for the application of ourmethods to the detection of intracellular forms ofmycobacteria. Our DNA extraction method has alsobeen successfully applied to an in vitro culturedspheroplast like organism recovered from CDtissue.29 We therefore believe that spheroplasts, ifpresent in CD tissue, would have been susceptible tothe extraction methods used. Nevertheless, theabsence of mycobacterial DNA at these levels ofsensitivities does not exclude a mycobacterialaetiology for CD because some mycobacterial infec-tions (lepromatous leprosy in man orM paratubercu-losis infections in ruminants) produce large numbersof bacilli, while others are associated with very fewundetectable organisms in tissues (tuberculosis andtuberculous leprosy). In this regard it is relevant toreport that attempts by us to detect M tuberculosisDNA in several human tuberculous cervical lymphnodes were unsuccessful (unpublished observations).A recent report has shown the presence of myco-

bacterial DNA in CD and control tissue derivedDNA using solution hybridisation with CD isolatedmycobacteria strain Linda total DNA as probe.24The thermal stability data for their DNA hybridsindicated that these sequences were neither identicalto each other nor to strain Linda and they concludedthat one mycobacterial species may not be specific-ally associated with CD. The levels of mycobacteriathey found ranged between one organism/two humancells to one/700 cells. The total absence of myco-bacterial DNA sequences observed in our presentstudy may reflect our slightly lower sensitivity rangeor the specificity of cloned probes which readily

distinguish a positive banding pattern from back-ground hybridisation. A previous study by us alsofailed to show mycobacterial DNA in mesentericlymph nodes from 21 cases of CD.-

Clearly even more sensitive techniques thanthose described here providing precise informationon the organisms detected are required. Gene ampli-fication techniques using the polymerase chainreaction have recently been described that mayenrich a specific DNA sequence more than a millionfold.3132 With the DNA sequence data for specificmycobacterial target DNA at present being obtainedin this laboratory, these methods combined withSouthern blotting and RFLP analysis should providecrucial information concerning the possible myco-bacterial aetiology of Crohn's disease.

This work was supported by grants from ActionResearch for The Crippled Child and The Anthonyand Elizabeth Mellows Charitable Trust to whom wewould like to express our sincere thanks.

References

1 Mayberry JF, Rhodes J. Epidemiological aspects ofCrohn's disease: a review of the literature. Gut 1984; 25:886-99.

2 Allan A. Inflammatory bowel disease: Aetiology andepidemiology. Curr Opin Gastroenterol 1985; 1: 461-7.

3 Mitchell DN, Rees RJW. Possible role of infectiousagents in Crohn's disease. Z Gastroenterol 1979; 17: 98-100.

4 Beeken WL. Transmissible agents in inflammatorybowel disease. Med Clin NAm 1980; 64: 1021-35.

5 Pena AS, Weterman IT, Booth CC, Strober W, eds.Recent advances in Crohn's disease. Developments inGastroenterology 1. London: Martinus-Nijhoff, 1981.

6 Patterson DSP, Allan WM. Chronic mycobacterialenteritis in ruminants as a model of Crohn's disease.Proc R Soc Med 1972; 65: 998-1001.

7 Chiodini RJ, van Kruiningen HJ, Merkal RS. Ruminantparatuberculosis (Johne's disease): the current statusand future prospects. Cornell Vet 1984; 74: 218-62.

8 Golde DW. Aetiology of regional enteritis. Lancet 1968;i: 1144-5.

9 Gitnick G. Is Crohn's disease a mycobacterial diseaseafter all? Dig Dis Sci 1984; 29: 1086-8.

10 Wilensky A, Moschcowitz E. Non specific granuloma ofthe small intestine. Am J Med Sci 1927; 173: 374-80.

11 Crohn BB, Ginsberg L, Oppenheimer GD. Regionalileitis. A pathological and clinical entity. JAMA 1932;99:1323-9.

12 Burnham WR, Lennard-Jones JE, Stanford JL, BirdRG. Mycobacteria as a possible cause of inflammatorybowel disease. Lancet 1978; ii: 693-6.

13 Belsheim MR, Darwish RZ, Watson WC, Schieven B.Bacterial form isolation from inflammatory boweldisease patients. Gastroenterology 1983; 85; 364-9.

14 Chiodini RJ, van Kruiningen HJ, Thayer WR, MerkalRS, Coutu JA. Possible role of mycobacteria in inflam-matory bowel disease. 1 An unclassified mycobacterium

on Decem

ber 21, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.29.9.1222 on 1 Septem

ber 1988. Dow

nloaded from

1228 Butcher, McFadden, and Hermon- Taylor

species isolated from patients with Crohn's disease. DigDis Sci 1984; 29: 1073-9.

15 Chiodini RJ, van Kruiningen HJ, Merkel RS, ThayerWR, Coutu JA. Characteristics of an unclassified myco-bacterium species isolated from patients with Crohn'sdisease. J Clin Microbiol 1984; 20: 966-71.

16 van Kruiningen HJ, Chiodini RJ, Thayer WR, CoutuJA, Merkal RS, Runnels PL. Experimental disease ininfant goats induced by a mycobacterium isolated from apatient with Crohn's disease. Dig Dis Sci 1986; 31: 1351-60.

17 Chiodini RJ, van Kruiningen HJ, Thayer WR, CouteauJA. Spheroplastic phase of mycobacteria from Crohn'sdisease. J Clin Microbiol 1986; 24: 357-63.

18 Chiodini RJ, van Kruiningen HJ, Thayer WR, CoutuJA. Differentiation of the mycobacterium sp. fromCrohn's disease by restriction fragment length poly-morphism of the ribosonal genes [Abstract]. Gastro-enterology 1986; 90: 1372.

19 McFadden JJ, Butcher PD, Chiodini RJ, Hermon-Taylor J. Determination of genome size and DNAhomology between an unclassified mycobacteriumspecies isolated from patients with Crohn's disease andother mycobacteria. J Gen Microbiol 1987; 133: 211-4.

20 McFadden JJ, Butcher PD, Chiodini RJ, Hermon-Taylor J. Crohn's disease-isolated mycobacteria areidentical to Mycobacterium paratuberculosis, as deter-mined by DNA probes that distinguish between Myco-bacterial species. J Clin Microbiol 1987; 25: 796-801.

21 McFadden JJ, Butcher PD, Thompson J, Chiodini R,Hermon-Taylor J. The use of DNA probes identifyingrestriction fragment length polymorphisms to examinethe Mycobacterium avium complex. Mol Microbiol1987; 1: 283-91.

22 Graham DY, Markesich DC, Yoshimura HH. Myco-bacteria and inflammatory bowel disease. Results ofculture. Gastroenterology 1987; 92: 436-42.

23 Collins JB, Beaman B, Arthur M, Gitnick G. Isolationof mycobacteria from intestinal tissues [Abstract].Gastroenterology 1986; 90: 177.

24 Yoshimura HH, Graham DY, Estes MK, Merkal RS.Investigation of association of mycobacteria with inflam-matory bowel disease by nucleic acid hybridisation.J Clin Microbiol 1987; 25: 45-51.

25 Patel R, Kuach JT, Mounts P. Isolation and restrictionendonuclease analysis of Mycobacterial DNA. J GenMicrobiol 1986; 132: 541-51.

26 Feinberg AP, Vogelstein B. A technique for radiolabel-ling DNA restriction fragments to high specific activity.Anal Biochem 1984; 137: 226-7.

27 Maniatis T, Fritsch EF, Sambrook J. Molecular cloning- A laboratory manual. New York: Cold Spring HarborLaboratory Publications, 1982.

28 Baess I. Determination and re-examination of genomesizes and base ratios on deoxyribonucleic acid fromMycobacteria. Acta Pathol Microbiol Immunol Scand1984; B92: 209-11.

29 McFadden JJ, Thompson J, Hull E, Hampson S,Stanford J, Hermon-Taylor J. The use of cloned DNAprobes to examine organisms isolated from Crohn'sdisease tissue. 1987 In: MacDermott, ed. Inflammatorybowel disease: current status and future approach.Amsterdam: Elsevier, 1988.

30 Butcher PD, McFadden JJ, Hermon-Taylor J. The useof cloned mycobacterial DNA probes for the detectionof Mycobacteria in Crohn's disease tissue. Biochem SocTrans 1987; 15: 547-9.

31 Saike RK, Scharf S, Faloona F, et al. Enzymaticamplification of beta globin genomic sequences andrestriction site analysis for diagnosis of sickle celldisease. Science 1985; 230: 1350-4.

32 Saike RK, Gelfand DH, Stoffel S, et al. Primer-directedenzymatic amplification of DNA with a thermostableDNA polymerase. Science 1988; 239: 487-91.

on Decem

ber 21, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.29.9.1222 on 1 Septem

ber 1988. Dow

nloaded from