Genetic relatedness analysis of Nocardia strains by random amplification polymorphic DNA: validation and applications

Res. Microbiol. 151 (2000) 263–270 © 2000 Éditions scientifiques et médicales Elsevier SAS. All rights reserved S0923250800001522/FLA

Genetic relatedness analysis of Nocardia strains by random amplification polymorphic DNA: validation and applications Frédéric Laurenta*,b, Frédérique Provost, Andrée Coubleb, Emmanuelle Casolib, Patrick Boironb a

b

Laboratoire de bactériologie, UMR CNRS 5558, Faculté de médecine Lyon-Sud, BP12, 69921 Oullins, France Laboratoire de mycologie fondamentale et appliquée aux biotechnologies industrielles, E.A 1655, Institut des sciences pharmaceutiques et biologiques de Lyon, 69373 Lyon, cedex 8, France (Submitted 27 April 1999; accepted 7 January 2000)

Abstract — Until now, no simple and rapid technique existed for epidemiological study of strains belonging to the Nocardia genus. The application of the arbitrarily primed PCR procedure to generate randomly amplified polymorphic DNA (RAPD) fingerprints for such analysis of Nocardia isolates was investigated. Fifty-one unrelated clinical isolates of N. asteroides were tested. Two conditions of RAPD using two different primers generated RAPD fingerprints that allowed the differentiation of all strains. The patterns were reproducible and discriminating. The results highlight the diversity of N. asteroides species and confirm that RAPD analysis is a highly valuable tool for studying the epidemiology of the Nocardia genus. Several examples describe the advantage of RAPD analysis for establishing the relationship between isolates from a given patient (long-term infections, coinfections) and from different patients (i.e. during an outbreak). In the future, this technique will help us to investigate the source of infection in cases of nosocomial transmission, to understand the outcome of nocardiosis, and to follow the evolution and acquisition of resistance to Nocardia strains. © 2000 Éditions scientifiques et médicales Elsevier SAS Nocardia / genotyping / RAPD / epidemiology

1. Introduction Nocardia species are ubiquitous in the environment. They are responsible for sporadic pulmonary diseases acquired by inhalation of spores with secondary localization in the central nervous system and subcutaneous tissues [9]. Several presumptive outbreaks of nocardiosis were observed in immunodepressed patients, especially in transplant units [1, 3, 6, 10, 12]. Some studies suggest, without demonstration,

* Correspondence and reprints Tel.: +33 (0)1 78 86 12 32; fax: +33 (0)1 78 86 32 80; [email protected] Abbreviations: PCR, polymerase chain reaction; PFGE, pulsed field gel electrophoresis; RAPD, random amplified polymorphic DNA; RFLP, restriction fragment length polymorphism; UPGMA, unweighted pair group method OD average

that infection could be transmissible as an airborne or a man-to-man infection [4, 6]. Such outbreaks require rapid infection control measures and diagnosis of nosocomial transmission and identification of the sources of the Nocardia strains are thus of crucial importance. In the same way, recurrences of nocardiosis had been suspected [13] but their studies could usually not be performed because of a lack of simple epidemiological tools. Indeed, phenotypic and serological characters have been shown to be of limited value for epidemiological investigation given their low discriminating power. DNA typing of Nocardia species will likely be a useful tool for distinguishing the different strains. But only a few and limited epidemiological molecular investigations have been reported: plasmid profiles [8], DNA-RFLP (restriction fragment length poly-

264

Laurent et al.

morphism, [5], PFGE (pulse field gel electrophoresis, [13], and RAPD (random amplified polymorphic DNA) [5] have been used without conclusive validation (for instance, such studies included less than five unrelated strains as control). Therefore, none of these techniques has definitively demonstrated their interest. Only a thorough evaluation of a significant number of related and unrelated isolates will definitively establish the usefulness of these techniques. Until now, this has not been achieved for the previously cited molecular techniques. Moreover, several of them require experience, time and heavy material. In this study, we evaluated RAPD analysis for epidemiological study of Nocardia isolates. We investigated the reproducibility, typeability, and discriminatory power of this technique using a large number of isolates. The interest of the method in the epidemiological setting was assessed by using related strains or undefined strains (i.e. strains for which epidemiological linkages were questionable).

2. Materials and methods 2.1. Bacterial strains

A total of 51 clinically unrelated strains of N. asteroides, i.e. collected from different geographic origins in France and around the world between 1992 and 1996 (including the type strain ATCC 19247), and isolated from different clinical sites (lung, brain abscess and skin) were analyzed (table I). All isolates were previously identified by conventional methods [2]. In order to validate RAPD for epidemiological purposes in the Nocardia genus, several related strains for which an epidemiological lineage was clearly established, were investigated (table II): i) three strains (strains N1, N2 and N3) isolated from cardiac transplant patients with pulmonary nocardiosis during a previously described outbreak; ii) two strains (N4 and N5) isolated from protected specimen brush and blood cultures of the same patient during an acute pulmonary infection. In contrast, we used two couples of strains for which epidemio-

logical relations had not been clearly established, to demonstrate the advantage of RAPD for the study of nocardiosis: i) two strains, N6 and N7, recovered from the same expectoration of a 36-year-old patient and showing different morphologies on plate agar, ii) two strains, N8 and N9, isolated within a 9-month period from the bronchoalveolar lavage of a 54-year-old patient presenting a persistent pulmonary infection. 2.2. DNA preparation and RAPD procedure

Bacteria were maintained on Bennett’s agar medium [7] at 25 °C. They were subcultured for 3–5 days at 34 °C on a sterile cellulose acetate membrane (Millipore, Bedford, MA) on blood agar (bioMérieux, Marcy-l’Etoile, France). A colony of each strain was suspended in 500 µL of Chelex solution (15% Chelext 100 resin (w/v) (sodium form, BioRad, Hercules, CA), 0.1% sodium dodecyl sulfate, 1% Nonidet P40 (v/w) and 1% Tween 80 (v/v)). The mixture was vortexed 90 s, then incubated for 30 min at 98 °C and centrifuged for 8 min at 10 000 g. The supernatant was recovered and 35-µL aliquots were used for amplification. The 50-µL DNA amplification reaction mixture contained 10 mM Tris-HCl pH 8.8, 50 mM KCl, 1.5 µM MgCl2, 0.1% Triton X-100, 200 µM of each dNTP, 2.5 U of Hi-Taq DNA polymerase (Bioprobe Systems, Montreuil-sous-Bois, France) and 1 or 0.1 µM of primer P1 (5’-CAATCGCCGT-3’) and P2 (5’CCGCCGACCGAG-3’), respectively. Amplified reactions were carried out in a DNA thermal cycler (Omnigene Hybrid, Teddington, UK) for 45 cycles consisting of 1 min at 94 °C, 1 min at 42 °C and 2 min at 72 °C (table III). The products were visualized by 1 mg/L ethidium bromide staining after electrophoresis in a 2% agarose gel. 2.3. Analysis of strain relatedness

The identity of RAPD patterns was defined on the basis of the similarity of numbers and matching positions of major bands. RAPD pattern normalizations were based on a 100-bp fragment and a RAPD pattern of a reference

Genetic relatedness analysis of Nocardia strains

265

Table I. Origin of the unrelated studied strains belonging to the N. asteroides species. Strain number

Reference

Geographic origina

Clinical site

National Reference Center 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

90.0679 90.0747 90.0923 90.0935 90.0964 90.1123 91.3075 91.3176 91.3509 92.0035 92.0124 92.0418 92.0648 93.0300 93.0303 93.0494 93.1026 93.1300 93.1388 93.1504 93.1587 93.1626 93.1678 93.1708 93.1798 94.0049 94.0061 94.0321 94.0322 94.0402 94.0416 94.0615 94.0652 94.0665 94.0802 94.0875 94.0934 94.0941 94.1039 94.1190 94.1198 94.1206 94.1281 94.1357 94.1519 94.1534 UC23 UC33 UC40 UC44 GUH-2

Bayonne (S.W.) Paris Tours (N.W.) Paris Saint-Etienne (S.E.) Paris Paris Chile Paris Switzerland Paris Algeria Vichy (N.E.) Nantes (N.W.) Nantes (N.W.) Kuwait Lyon (S.E.) Spain Paris Paris Martinique Nantes (N.W.) Arles (S.E.) Nice (S.E.) Paris Bordeaux (S.W.) Colmar (N.E.) Grenoble (S.E.) Paris Paris Le Havre (N.W.) Strasbourg (N.E.) Périgueux (S.W.) Lorient (N. W.) Angers (N.W.) Rodez (S.W.) Strasbourg (N. E.) Montpellier (S.W.) Nancy (N. E.) Lyon (S.E.) Bourges (N.W.) Toulouse (S.W.) Toulouse (S.W.) Paris Lyon (S.E.) Nantes (N.W.) United States United States United States United States United States

Lung Lung Blood Brain Skin Skin Brain Skin Brain Lung Skin Lung Blood Lung Lung Brain Lung Brain Lung Blood Skin Lung Lung Skin Brain Skin Lung Lung Skin Brain Lung Lung Blood Lung Blood Lung Brain Skin Blood Lung Blood Lung Lung Lung Blood Lung Lung Brain Lung Lung Skin

RAPD pattern with primer P1

P2

Pattern with combination of results with primers P1 and P2

A1 A2 A3b A4 A5 A6 A7 A8 A9 A10 A11 A12 A13 A14 A15 A16 A17 A18 A19 A20 A21 A22 A23 A24 A25 A26 A27 A28 A29 A30 None A31 A32 A18 A33 None A34 A35 A36 A37 A38 A39 A40 A41 A42 A43 A44 A45 A3 A46 A47

B1 B2 B3 B4 B5 B6 B7 B8 B9 B10 B11 B12 B13 B14 B15 B16 B17 B18 B19 B20 B21 B22 B23 None B24 B25 B26 B27 B28 B29 B30 B31 B32 B33 B34 None B35 B36 B37 B38 B39 B40 B41 B42 B43 B44 B45 B46 B45 B47 B48

W1 W2 W3 W4 W5 W6 W7 W8 W9 W10 W11 W12 W13 W14 W15 W16 W17 W18 W19 W20 W21 W22 W23 W24 W25 W26 W27 W28 W29 W30 W31 W32 W33 W34 W35 None W36 W37 W38 W39 W40 W41 W42 W43 W44 W45 W46 W47 W48 W49 W50

a Geographic origin: France was divided into five geographic areas. N.W., northwest; N.E., northeast; S.E., southeast; S.W., southwest; Paris. b Strains showing no amplified fragment or identical patterns are indicated in bold face.

266

Laurent et al.

Table II. Origin of the related and undefined studied strains. No. N1 N2 N3 N4 N5 N6 N7 N8 N9

Geographic origin

Genus

Caen Caen Caen Paris Paris Paris Paris Toulouse Toulouse

Nocardia Nocardia Nocardia Nocardia Nocardia Nocardia Nocardia Nocardia Nocardia

Species farcinica farcinica farcinica asteroides asteroides asteroides asteroides asteroides asteroides

Clinical site RAPD pattern with primer Lung Lung Lung Lung Blood Lung Lung Lung Lung

strain included in each experiment. Two strains having at least one difference according to these major bands (difference of size greater than 5%) were considered as different. The small differences in faint bands were considered only above two and the strains then classified as different. Comparisons between the fingerprints were performed with the Taxotront package (Taxolab, Institut Pasteur, Paris, France). The unweighted pair group method OD averages (UPGMAs) was used for cluster analysis and led to establishment of normalized drafts and dendrograms for each of the amplification conditions [11].

3. Results 3.1. Reproducibility

To evaluate the reproducibility of the method, five isolates (numbers 1–5 in table I) were tested five times by preparing DNA extracts on different days after subcultures and amplification on different days. Identical banding patterns were

P1

P2

A48 A48 A48 A49 A49 A50 A50 A51 A52

B49 B49 B49 B50 B50 B51 B51 B52 B53

Pattern with combination of results with primer P1 and P2 W51 W51 W51 W52 W52 W53 W53 W54 W55

revealed for each isolate with the two amplification conditions when we accepted a 5% deviation in size. Moreover, the type strain N. asteroides ATCC 19247 was included in each RAPD assay as an external standard. It always showed the same pattern. 3.2. RAPD profiles

The low stringency of the PCR conditions enabled the amplification of one to eleven fragments. Depending on the primer, fragments ranged in size from 100 to 2 000 bp. For the rare isolates that gave a pattern with a unique band, we had systematically confirmed the results by repeating the RAPD analysis (table I). With the primer P1, two strains (94.0412 and 94.0875) showed no amplified band and two couples of strains (90.0923/UC40 and 93.1300/94.0665) showed similar profiles. With the primer P2, a unique strain (94.0875) showed no amplified band and strains UC23 and UC40 showed an identical pattern (table I). It must be noted that strains with identical profiles according to one of the amplification conditions revealed only

Table III. Conditions and results of performance criteria for the amplifications used. T = Nt/N where T is the typeability of the technique, Nt the number of isolates assigned a type and N the number of isolates tested; D = 1 − 兺 nj共 nj − 1 兲/N共 N − 1 兲where D is the index of discriminatory power, N the number of unrelated strains tested and nj the number of strains belonging to the j type. Primers Concentration (µM) T° of annealing (°C) Typeability (T) Discriminatory power (D)

P1

P2

Combined

1 42 0.94 0.998

0.1 42 0.96 0.999

– – 0.98 0.999

Genetic relatedness analysis of Nocardia strains

267

one or two bands. Such results cannot be considered as informative and specific. Therefore, these same strains were always easily differentiated when we considered their pattern using the second primer. 3.3. Typeability

According to our data, typeability (T = Nt/N where T is the typeability of the technique, Nt the number of isolates assigned a type and N the number of isolates tested) of the RAPD technique was high (0.96 for each tested primer) and reached 0.98 when we combined the results obtained with the two primers (one strain nontypeable: no amplification profile (94.0875)) (table I). 3.4. Discriminatory power

Discriminatory power (D = 1 − 兺 nj共 nj − 1 兲/ N共 N − 1) where D is the index of discriminatory power, N the number of unrelated strains tested and nj the number of strains belonging to the j type) was high and overtook 0.998, 0.999 and 0.999, respectively, with primer P1, primer P2 and combined results (table III). 3.5. Dendrograms

With the Taxotront package using the average linked method of clustering (UPGMA), two dendrograms for each primer were built. They enabled the evaluation of the genetic relationships among the isolates. Genetic differences were expressed with clusters (figure 1).

Figure 1. Dendrogram constructed from 51 unrelated N. asteroides isolates according to the fingerprints using primer P2. The tree was generated from the distance matrix by UPGMA.

3.5.1. Unrelated strains

RAPD analysis of these strains showed that they were related neither to the geographic origins nor to the clinical origin (lung, skin or brain abscesses). Instead, there was no common arm when we considered the two trees of the two generated dendrograms: isolates which were clustered according to one of the primers were not related when we considered the second primer, indicating high variability of the N. asteroides strains.

3.5.2. Related strains

RAPD profiles of the three strains (N1, N2, N3) isolated during a documented outbreak in a cardiac transplant unit were identical whatever the two conditions of amplification, confirming the clonality of the isolates. The two strains (N4 and N5) isolated in BAL liquid and a blood culture from the same patient presenting acute pulmonary disease supported identical arrays of amplified fragments (figure 2).

268

Laurent et al.

Figure 2. RAPD fingerprints from Nocardia strains obtained using primer P2. Reference indicated above each lane refers to isolate identification in table II. M is a 100-bp DNA ladder.

3.5.3. Undefined strains

The two isolates (N6 and N7) recovered from the same expectoration culture but showing two different profiles (figure 2) correlated well with the two different morphologies observed on the agar plates. In contrast, the two strains (N8, N9) isolated within a period of 9 months from the same patient revealed the same profile (figure 2), indicating persistence of the same strain.

4. Discussion In the last 10 years, a new PCR-based method of DNA fingerprinting, known as the RAPD assay, has been described for typing many bacterial species [15, 16]. It should be considered as an attractive option for epidemiological study of Nocardia: i) knowledge of genetic sequences are not necessary (arbitrary primer); ii) the process is very rapid; excluding chromosomal DNA, it can be finished in 1 day compared to 2–3 days for RFLP, plasmid profiles or PFGE; iii) it is less cumbersome and less expensive than many molecular techniques. The resolving power of epidemiological typing of organisms has been expanded by a

molecular analysis of microbial DNA. In contrast to the majority of conventional typing methods based on species-restricted variation of antigenic, metabolic or other phenotypic determinants, similar strategies of DNA analysis can be applied to any bacteria. But new typing methods are often applied without critical evaluation of their performance characteristics. In this study, we have evaluated our RAPD protocol as an epidemiological typing tool for bacteria belonging to the genus Nocardia as recommended by the consensus guidelines of the European Study Group on Epidemiological Markers of the European Society for Clinical Microbiology and Infectious Diseases [14]. We have studied a large number (n = 51) of Nocardia strains. Our results indicated that the RAPD analysis presented good reproducibility and was a suitable method for distinguishing between unrelated strains when the two RAPD conditions were used. The index of discrimination was nearly 100% when results of the two amplification conditions used were combined. In contrast, strains for which epidemiological linkages were highly suspected (N1 to N3 and N4 to N5) presented identical RAPD profiles. These results confirm the capacity of the RAPD technique to differentiate clonal and nonclonal isolates. The two dendrograms obtained in the present study with unrelated strains show no association among isolates according to geographic origin or anatomic sites. Similarly, the lack of correlation between the two phylogenetic trees demonstrated random distribution of strains, indicating a great deal of genetic heterogeneity at the species level despite phenotypic homogeneity. These results revealed that the number of clones involved in transmission and infection is very high. The relatively extensive genetic diversity might be due to the fact that Nocardia have a world-wide distribution in soil and that selection pressure impact on the organism in various environmental niches is weak. The major application of strain characterization is the confirmation that isolates of Nocardia are identical in clusters of cases. This is not simply of academic interest, as demonstrated by

Genetic relatedness analysis of Nocardia strains

several presumptive outbreaks previously described [4, 6, 15]. Such outbreaks were sometimes staggered over several months and were then difficult to recognize and confirm. They were saddled with high mortality (about 50%) and required rapid infection control measures including ward closure (for more than 1 month, on occasion) followed by thorough cleaning and formaldehyde fumigation [6, 10]. The human and economic cost is particularly high. Therefore, a rapid assessment of nosocomial acquisition of nocardiosis occurring in cluster of cases is necessary. RAPD analysis is an appropriate response. RAPD profiles of the three N. farcinica isolates (N1, N2, N3) were identical and different from one another. These results confirmed the clonality of the strains in these nosocomial presumptive cases. The RAPD protocol we have developed could enable the identification of reservoir(s) of epidemic clone(s) in the population (colonization in the upper respiratory tract has been described) and/or in the environment. Moreover, it could be of use in epidemiological studies by generating hypotheses concerning the identification of the source(s) of contamination and the vehicles of transmission. Finally, it could guide health resources for appropriate measures to control the outbreak and enable evaluation of their efficacy [10]. Another application of RAPD is the study of the pathogenesis and natural history of nocardiosis, which remains to be established. RAPD could enable us to better understand persistent nocardiosis. The RAPD profiles could highlight the relative role of endogenous reactivating chemotherapy failure, exogenous reinfection and/or multiple infection as the cause of recurrent or long-term persistent infection, which are sometimes reported. For example, the two identical band patterns of N8 and N9 strains confirmed, in this case, the persistence of the same strain despite antibiotherapy. On the contrary, the two different patterns obtained for strains N6 and N7 demonstrate that coinfections with different Nocardia strains are possible for a given patient.

269

Our results enable validation of RAPD as a simple and rapid epidemiological typing tool for Nocardia strains. This technique that we have has revealed the diversity of Nocardia strains and the lack of relatedness of strains in terms of geographic origins and clinical sites. Its high discriminative capacity could be useful for epidemiological study, especially for investigating the sources of outbreaks, the relatedness of isolates recovered from different patients, or of serial isolates from the same patient, and modifications in susceptibility profiles. Résumé — Étude épidémiologique des souches appartenant au genre Nocardia par RAPD : validation et applications. La technique de RAPD (random amplified polymorphic DNA) a été évaluée comme outil moléculaire pour l’étude épidémiologique des souches appartenant au genre Nocardia. Cette technique a démontré, dans les conditions du protocole que nous avons développées, une reproductibilité, une typabilité et un pouvoir discriminant excellents. Les résultats mettent en évidence la diversité des souches cliniques rencontrées. Comme nous l’avons montré à l’aide de quelques exemples, cette technique permet d’étudier les liens éventuels existant entre les souches isolées de différents patients (bouffée épidémique), de comprendre, de suivre et d’étudier les formes de nocardiose chronique ou les cas d’infections multiples. Elle devrait permettre à l’avenir d’étudier les voies de transmission ainsi que l’évolution et l’acquisition de résistance chez ces bactéries. © 2000 Éditions scientifiques et médicales Elsevier SAS Nocardia / RAPD / génotypie / épidémiologie

References [1] Baddour L.M., Baselski V.S., Herr M.J., Christensen G.D., Bisno A.L., Nocardiosis in recipients of renal transplants: evidence for nosocomial acquisition, Am. J. Inf. Control 14 (1986) 214–219. [2] Boiron P., Provost F., Dupont B., Laboratory Methods for the Diagnosis of Nocardiosis, Institut Pasteur Editions, Paris, 1993. [3] Chapman S.W., Wilson J.P., Nocardiosis in transplant recipients, Sem. Resp. Infect. 5 (1990) 74–79. [4] Cox F., Hughes W.T., Contagious and other aspects of nocardiosis in the compromised host, Pediatrics 55 (1975) 135–138. [5] Exmelin L., Malbruny B., Vergnaud M., Provost F., Boiron P., Morel C., Molecular study of nosocomial nocardiosis outbreak involving heart transplant recipients, J. Clin. Microbiol. 34 (1996) 1014–1016. [6] Houang E.T., Lovett I.S., Thompson F.D., Harrison A.R., Joekes A.M., Goodfellow M., Nocardia asteroides infection: a transmissible disease, J. Hosp. Infect. 1 (1980) 31–40.

270

Laurent et al.

[7] Jones K.L., Fresh isolates of actinomycetes in which the presence of sporogenous aerial mycelia is a fluctuating characteristic, J. Bact. 57 (1949) 141–145. [8] Jonsson S., Wallace Jr R.J., Hull S.I., Musher D.M., Recurrent Nocardia pneumonia in an adult with chronic granulomatous disease, Am. Rev. Resp. Dis. 133 (1986) 932–934. [9] McNeil M.M., Brown J.M., The medically important aerobic actinomycetes: epidemiology and microbiology, Clin. Microbiol. Rev. 7 (1994) 357–417. [10] Sahathevan M., Harvey F.A.H., Forbes G., O’Grady J., Gimson A., Bragman S., Jensen R., Philpott-Howard J., Williams R., Casewell M.W., Epidemiology, bacteriology and control of an outbreak of Nocardia asteroides infection on a liver unit, J. Hosp. Infect. 18 (1991) 473–480. [11] Sneath P.H.A., Sokal R.R., Numerical Taxonomy, W. H. Freeman Editions, San Francisco, 1973. [12] Stevens D.A., Pier A.C., Beaman B.L., Morozumi P.A., Lovett I.S., Houang E.T., Laboratory evaluation of an outbreak of nocardiosis in immunocompromised hosts, Am. J. Med. 71 (1981) 928–934.

[13] Steingrube V.A., Wallace Jr R.J., Brown B.A., Pang Y., Zeluff B., Steele L.C., Zhang Y., Acquired resistance of Nocardia brasiliensis to clavulanic acid related to a change in b-lactamase following therapy with amoxicillin-clavulanic acid, Antimicrob. Agent Chemother. 35 (1991) 524–528. [14] Struelens M.J. and the members of the European study group on epidemiological markers of the European Society for Clinical Microbiology and Infectious Diseases, Consensus guidelines for appropriate use and evaluation of microbial epidemiologic typing systems, Clin. Microbiol. Infect. 3 (1995) 121–130. [15] Welsh J., McClelland M., Fingerprinting genomes using PCR with arbitrary primers, Nucleic Acid Res. 18 (1990) 7213–7218. [16] Williams J.G.K., Kubelik A.R., Livak K.J., Rafalski J.A., Tingey S.V., DNA polymorphism amplified by arbitrary primers are useful as genetic markers, Nucleic Acid Res. 18 (1990) 6531–6535.