New Nocardia Taxon among Isolates of Nocardia brasiliensis ...

4 downloads 83 Views 342KB Size Report
Jan 24, 1995 - ... JR.,1* BARBARA A. BROWN,1 ZETA BLACKLOCK,2 ROBERT ...... Wallace, R. J., Jr., E. J. Septimus, T. W. Williams, Jr., R. H. Conklin, T. K..
JOURNAL OF CLINICAL MICROBIOLOGY, June 1995, p. 1528–1533 0095-1137/95/$04.0010 Copyright q 1995, American Society for Microbiology

Vol. 33, No. 6

New Nocardia Taxon among Isolates of Nocardia brasiliensis Associated with Invasive Disease RICHARD J. WALLACE, JR.,1* BARBARA A. BROWN,1 ZETA BLACKLOCK,2 ROBERT ULRICH,1 KEN JOST,3 JUNE M. BROWN,4 MICHAEL M. MCNEIL,4 GRACE ONYI,1 VINCENT A. STEINGRUBE,1 AND JEREMY GIBSON1 Nocardia/Mycobacteria Research Laboratory, Department of Microbiology, University of Texas Health Center, Tyler,1 and Texas Department of Health, Austin,3 Texas; Mycobacteriology Laboratory, Queensland State Health Laboratories, Brisbane, Australia2; and Emerging Bacterial and Mycotic Diseases Branch, Division of Bacterial and Mycotic Diseases, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia4 Received 28 October 1994/Returned for modification 24 January 1995/Accepted 18 March 1995

Nocardia brasiliensis, the second most frequently isolated aerobic actinomycete in the clinical laboratory, is usually associated with localized cutaneous infections. However, 22% of 238 N. brasiliensis isolates from the United States and 12% of 66 isolates from Queensland, Australia, which had been collected over a 17-year period, were associated with extracutaneous and/or disseminated diseases. Of the 62 invasive isolates, 37 (60%) were susceptible to ciprofloxacin and/or were susceptible to clarithromycin and resistant to minocycline, compared with only 6 (3%) of 242 localized cutaneous isolates. The 43 isolates with this susceptibility pattern appeared to define a new taxon. They were similar to Nocardia asteroides complex isolates clinically in proportions from persons with pulmonary (70%), central nervous system (23%), and/or disseminated diseases (37%) in the setting of corticosteroids (74%) or AIDS (14%). This putative new taxon differed from N. brasiliensis in the hydrolysis of adenine (92 versus 4%), b-lactamase patterns on isoelectric focusing, and the presence of two early mycolic acid-ester peaks by high-performance liquid chromatography. Restriction analysis of a 439-bp fragment of the 65-kDa heat shock protein gene revealed that N. brasiliensis and the new taxon had different restriction patterns with 8 of the 11 enzymes tested. Screening of invasive isolates of N. brasiliensis for susceptibility to ciprofloxacin will identify most isolates of the new taxon, which likely represents a new Nocardia species. vealed that these isolates represent a putative new invasive taxon that masquerades in its hydrolysis patterns as N. brasiliensis. The purpose of this study was to present the clinical and laboratory features of this new taxon.

Recent work at the University of Texas Health Center at Tyler (UTHCT) has focused on the use of antimicrobial susceptibility patterns in the genus Nocardia to enhance taxonomic grouping. Isolates of Nocardia asteroides complex are especially heterogeneous taxonomically, but clinical efforts to define specific subgroups have been frustrated by the absence of readily available reproducible testing methods. Carbohydrate utilization, chromatography, including high-performance liquid chromatography (HPLC) (6), and other routine test methods have all failed to improve on the standard clinical hydrolysis tests of casein, tyrosine, hypoxanthine, and xanthine. With hydrolysis testing, two common human pathogenic Nocardia spp., N. asteroides complex and N. brasiliensis, are easily separated. Other less frequently encountered species, including N. carnea, N. transvalensis, and perhaps N. otitidiscaviarum (formerly N. caviae) (10, 13), are not separated as well by this technique. Using antimicrobial susceptibility patterns, we were recently able to facilitate the clinical recognition of the previously described but hard to identify species N. nova (20) and N. farcinica (25) within the N. asteroides complex. Isolates of N. brasiliensis have not been considered heterogeneous taxonomically (2), although most studies have not included large numbers of isolates (10). During routine susceptibility testing on N. brasiliensis, we noted that some isolates that were clustered among those from pulmonary sources were unusually resistant to minocycline. Detailed assessment re-

MATERIALS AND METHODS Organisms. Clinical isolates were obtained from four laboratories. Isolates of Nocardia spp. that had been submitted to the Nocardia/Mycobacteria Research Laboratory of UTHCT for susceptibility testing from 1982 to August of 1994 were screened for isolates of N. brasiliensis with identified tissue sources and available susceptibility results. Isolates that had been submitted to one of us (R.J.W.) at Baylor College of Medicine (Houston, Tex.) between 1976 and 1982 were screened in a similar manner. Isolates that had been submitted to the Centers for Disease Control and Prevention, Atlanta, Ga., between 1986 and 1994 and Nocardia isolates that had been recovered from the State Health Laboratory, Queensland, Australia, between 1987 and early 1994 were also screened. The strain N. brasiliensis ATCC 19296T was obtained from the American Type Culture Collection (ATCC). Routine identification and susceptibility testing were done at the time of submission, while other testing was done on isolates that had been stored in 15% glycerol in Trypticase soy broth at 2708C. Clinical data. Routine clinical information, including culture source, was obtained from the referring laboratory at the time the isolate was submitted. Efforts were made to obtain detailed information on patients with extracutaneous disease and isolates from the new group. Identification. Isolates from the United States were identified as N. brasiliensis by the referring laboratory, most by the Centers for Disease Control and Prevention and the Texas Department of Health, by using standard criteria (10, 12, 13), including typical Gram stain and colonial morphology, resistance to lysozyme, and hydrolysis of casein, hypoxanthine, and tyrosine, not xanthine. Confirmation of species identification in the UTHCT laboratory was made by antibiotic susceptibility patterns, specifically resistance to kanamycin and cefamandole by agar disk diffusion (26). Queensland isolates were identified in the Queensland State Health Laboratories by the methods of Mishra et al. (13). Thirty-four isolates of the new taxon as well as 20 selected isolates of N. brasiliensis sensu stricto were retested in the Queensland laboratory for these features, as well as acid production from 20 carbohydrates, decomposition of

* Corresponding author. Mailing address: Department of Microbiology, The University of Texas Health Center, P.O. Box 2003, Tyler, TX 75710. Phone: (903) 877-7680. Fax: (903) 877-7652. 1528

NEW NOCARDIA TAXON

VOL. 33, 1995

1529

TABLE 1. Incidence of new taxon isolates among isolates that had been associated with cutaneous and invasive diseases and identified as N. brasiliensis Cutaneous isolatesa

Invasive isolatesa

Laboratory and/or location

Collection yr

Total no. of isolates

Total

New taxon (%)

Total

New taxon (%)

Texas CDC, Georgia Queensland, Australia

1976–1994 1986–1994 1987–1994

205 33 66

167 17 58

5 (3) 0 (0) 1 (2)

38 16 8

26 (68) 6 (38) 5 (63)

304

242

6 (3)

62

37 (60)

Total a

Isolates from sites other than skin or soft tissue (e.g., lung or brain) or from multiple noncontiguous sites, including skin (disseminated).

urea and adenine, nitrate reduction (24 h), 14-day arylsulfatase activity, and growth for 2 weeks at 10, 37, and 438C. Susceptibility testing. Isolates were tested by disk diffusion at both the UTHCT and Queensland laboratories for susceptibility to minocycline (30 mg), ciprofloxacin (5 mg), and trimethoprim-sulfamethoxazole (1.25/23.75 mg, where 1.25 and 23.75 mg are the amounts of trimethoprim and sulfamethoxazole, respectively) by using Mueller-Hinton agar and 72 h of incubation at 358C in room air (22, 23). MICs were determined at the UTHCT and Centers for Disease Control and Prevention (CDC) laboratories by broth microdilution in Mueller-Hinton broth (23, 24). Control strains were used with each set of plates. The drugs tested and their resistance breakpoints were amikacin (.16 mg/ml), minocycline (.8 mg/ml), cefotaxime and ceftriaxone (.32 mg/ml), amoxicillinclavulanic acid (2:1 ratio) (.8/4 mg/ml), ciprofloxacin (.2 mg/ml), clarithromycin (.4 mg/ml), and imipenem (.8 mg/ml). Intermediate values were defined for ciprofloxacin (2 mg/ml), minocycline, (8 mg/ml), and amoxicillin-clavulanic acid (8/4 mg/ml). These MIC breakpoints are similar to those of the National Committee for Clinical Laboratory Standards for aerobic organisms (14) and have been published previously (22, 23). New taxon definition. Isolates that had been identified by standard hydrolysis tests as N. brasiliensis were considered to belong to the new taxon if they met one of two susceptibility characteristics, (i) susceptibility to ciprofloxacin (MIC, #0.5 mg/ml; or disk zone size, $35-mm diameter) or (ii) high-level susceptibility to clarithromycin (MIC, #0.5 mg/ml) and intermediate susceptibility or resistance to minocycline (MIC, $8 mg/ml; or disk zone size, #15-mm diameter). Isolates that had been identified as N. brasiliensis and did not belong to the new taxon were referred to as N. brasiliensis sensu stricto. HPLC. Mycolic acid–p-bromophenacyl ester patterns were determined for isolates of the new taxon and selected invasive and cutaneous isolates of N. brasiliensis sensu stricto by HPLC (5, 6), as modified by Jost and Dunbar (7). b-Lactamase patterns by isoelectric focusing. Selected organisms were grown in Mueller-Hinton broth supplemented with 10% oleic acid, albumin, and dextrose and washed three times with acetone, and the enzyme was eluted overnight at 48C (18, 21). b-Lactamase preparations were focused on an LKB Multiphor 2117 with commercial Ampholine polyacrylamide gels according to manufacturer’s instructions and as previously described (18, 21). b-Lactamase bands were visualized with nitrocefin. PCR-restriction fragment length polymorphism (RFLP) analysis of a 65-kDa heat shock protein gene sequence. Selected isolates of Nocardia spp. were grown in Mueller-Hinton broth, harvested, suspended in 10 mM Tris–1 mM EDTA (pH 8.0), and heat inactivated at 808C for 30 min. Cells were disrupted in a Mini-Bead Beater-8 (Biospec Products, Bartlesville, Okla.) for 5 min in the presence of 0.1to 0.15-mm-diameter zirconium beads. Cell supernatants were clarified by two centrifugations. Each sample was then subjected to PCR for amplification of a 439-bp fragment of the 65-kDa heat shock protein gene with primers TB11 (59-ACCAACGAT GGTGTGTCCAT) and TB12 (59-CTTGTCGAACCGCATACCCT) as described by Telenti et al. (19), with Taq DNA polymerase. The 439-bp product was digested with 11 restriction enzymes (BstEII, AciI, CfoI, HaeIII, BsaHI, MspI, SstII, BanII, BstNI, HinfI, and NaeI) (New England Biolabs, Beverly, Mass.) according to manufacturer’s instructions and then electrophoresed on 3% Metaphore agarose (FMC Bioproducts, Rockland, Maine). Fragment sizes (in base pairs) were estimated on a computerized BioImage system (Millipore, Bedford, Mass.) with a 100-bp commercial ladder (Life Technologies, Grand Island, N.Y.) and a pGEM base pair ladder (Promega, Madison, Wis.). Only fragments of 100 bp or greater were recorded.

RESULTS Organisms. During this 17-year study, approximately 1,100 Nocardia isolates were submitted to the UTHCT laboratory. Of these, 225 were N. brasiliensis; tissue sources and susceptibility data were available for 205. Thirty-eight (19%) of these 205 isolates were classified as extracutaneous on the basis of

recovery from sites other than skin or adjacent soft tissue (i.e., brain, cerebrospinal fluid, mediastinum, kidney, blood, bone marrow, pleural fluid, sputum, bronchial washing, or lung) and/or disseminated (defined as recovery from blood, brain, or two or more noncontiguous sites, including skin). Of these 38 isolates, 30 (79%) involved a lung, 9 (24%) involved the central nervous system (CNS), and/or 16 (42%) were disseminated. A higher incidence of invasive disease was associated with CDC isolates, with a lower incidence among Queensland isolates (Table 1). These new taxon isolates were unevenly divided between extracutaneous and cutaneous isolates. Overall, 30 of 43 new taxon isolates (70%) involved a lung, 10 of 43 (23%) involved the CNS, and/or 16 of 43 (37%) resulted in disseminated disease (Table 2). According to the new taxon identification criteria based on susceptibility results, 37 of 238 (16%) isolates from the two United States laboratories and 6 of 66 (9%) Australian isolates that had been identified by standard methods as N. brasiliensis belonged to the new group (Table 1). Two of these new taxon isolates were submitted to the ATCC. ATCC 51511 had been recovered from the skin and liver of a patient with chronic leukemia during a fatal case of nocardiosis, while ATCC 51512 had been recovered from the leg, lung, and brain of a patient with ulcerative colitis. The 34 United States isolates with available data for geo-

TABLE 2. Clinical characteristics of 43 patients with isolates of the new taxon Clinical characteristic

Underlying diseases Autoimmune disorders Malignancy AIDS Organ transplantation Alcoholism Miscellaneousa None Unknown Total Sites of involvement Lungs Brain and/or meninges Skin and/or soft tissue Joint Miscellaneous Disseminated a

No. of cases (%)

No. of patients On steroids

With available steroid history

9 9 5 5 3 3 1 8

9 7 1 5 0 1 0 0

9 7 5 5 2 2 1 0

43

23

31

30 (70) 10 (23) 13 (30) 3 (7) 2 (5) 16 (37)

Diabetes, chronic lung disease, and immune suppression not specified.

1530

WALLACE ET AL.

graphic sources came from nine states. The most common sources were the states of Florida (15 isolates), Texas (6 isolates), and Georgia (3 isolates). Interestingly, 30 of 34 isolates with identified sources (88%) were from Southern coastal states. Clinical data. Clinical histories were obtained for 35 of 43 (81%) patients who had been infected with isolates of the new taxon (Table 2). Among these patients, 26% had autoimmune disorders, 26% had malignancies, 14% had organ transplants, and 14% had AIDS. Only 1 patient (2%) had no underlying disease; 23 of 31 (74%) patients had detailed histories of receiving corticosteroids. By excluding patients with chronic alcoholism or human immunodeficiency virus infection, 22 of 24 (92%) remaining patients had received corticosteroids. Identification. Isolates of the new taxon had typical Nocardia colony morphology, with all of them producing orange pigmentation in 10 days. They were gram positive and acid fast or partially acid fast by the modified Kinyoun method (1% sulfuric acid decolorization), but they were not acid fast by the standard Kinyoun method. The new taxon isolates were similar to N. brasiliensis sensu stricto isolates in urease activity (24 of 24), resistance to lysozyme (36 of 36), and hydrolysis of casein (36 of 36), hypoxanthine (32 of 35), and tyrosine (33 of 34), not xanthine (0 of 36). All isolates were resistant to kanamycin and cefamandole by disk diffusion, features that had previously been identified as typical of N. brasiliensis (26). Thirty-four isolates of the new taxon and 20 isolates of N. brasiliensis sensu stricto produced similar patterns of acid production from carbohydrates. Isolates of the new taxon were positive for inositol (91%), D-mannitol (91%), D-mannose (94%), D-glucose (94%), glycerol (100%), D-trehalose (94%), D-galactose (91%), and citrate (96%), while they were negative for adonitol (0%), L-arabinose (0%), cellobiose (3%), erythritol (0%), lactose (11%), maltose (0%), melibiose (0%), melezitose (0%), raffinose (0%), L-rhamnose (3%), D-sorbitol (3%), and xylose (6%). These isolates were negative for arylsulfatase and grew at 378C, but not at 10 or 438C, after 2 weeks. Differences between isolates of the new taxon and isolates of N. brasiliensis sensu stricto were observed for adenine hydrolysis and nitrate reduction. Twenty-eight of 33 (85%) new taxon isolates hydrolyzed adenine, compared with only 2 of 47 (4%) isolates of N. brasiliensis sensu stricto, while 5 of 28 (19%) of the former were nitrate positive, compared with 15 of 18 (83%) of the latter. Susceptibility testing. Numerous differences in susceptibility were observed between isolates of the new taxon and isolates of N. brasiliensis sensu stricto. Thirty-eight of 40 isolates (95%) of the new taxon were susceptible to ciprofloxacin (MIC, #0.5 mg/ml), with a mode of 0.25 mg/ml, while the ciprofloxacin MICs for 144 cutaneous and extracutaneous isolates of N. brasiliensis sensu stricto were $2 mg/ml, with a mode of 4.0 mg/ml (Table 3). Susceptible isolates had disk zone sizes of 35 to 60 mm in diameter. Among the 66 Australian isolates tested against ciprofloxacin, 6 (9%) had similar zone sizes and were considered to be susceptible. A similar bimodal distribution of susceptibilities was seen with minocycline. By disk diffusion, 31 of 37 isolates (84% of the new subgroup) had #15-mm-diameter zones, compared with 4 of 160 (2.5%) cutaneous and extracutaneous isolates of N. brasiliensis sensu stricto. By MIC determination, the minocycline MICs for 22 of 31 (73%) isolates of the new taxon were .8 mg/ml, compared with only 1 of 149 (0.7%) N. brasiliensis sensu stricto isolates (Table 3). The distribution of susceptibility to clarithromycin was similar to that of ciprofloxacin. Twenty-nine of 32 (91%) isolates

J. CLIN. MICROBIOL. TABLE 3. Antimicrobial susceptibilities of isolates of the new taxon compared with those of two clinical groups of N. brasiliensis sensu stricto isolates No. of isolates within parameter/no. of isolates testedb (%) Antimicrobial agent and susceptibility parametera

N. brasiliensis sensu stricto New taxon

Ciprofloxacin MIC # 0.5 mg/ml (S) MIC $ 2.0 mg/ml (I, R) Zone $ 35 mm (S) Minocycline MIC # 4 mg/ml (S) MIC 5 8 mg/ml (I) MIC . 8 mg/ml (R) Zone # 15 mm (R) Clarithromycin MIC # 0.5 mg/ml (S) MIC 5 1.0–4.0 mg/ml (I) MIC $ 8.0 mg/ml (R) Trimethoprim-sulfamethoxazole MIC # 32 mg/ml (S) Zone , 20 mm (R) Zone $ 20 mm (S) Amoxicillin-clavulanic acid MIC # 4/2 mg/ml (S) MIC 5 8/2 mg/ml (I) MIC . 8/2 mg/ml (R) Cefotaxime MIC # 32 mg/ml (S) MIC . 32 mg/ml (R) Ceftriaxone MIC # 32 mg/ml (S) MIC . 32 mg/ml (R) Amikacin MIC # 2 mg/ml (S) MIC 5 4–16 mg/ml (S) MIC . 16 mg/ml (R)

Extracutaneous

Cutaneous

38/40 (95) 2/40 (5) 35/37 (95)

0/21 (0) 21/21 (100) 0/13 (0)

0/123 (0) 123/123 (100) 0/164 (0)

4/31 (13) 5/31 (13) 22/31 (73) 31/37 (84)

8/11 (73) 3/11 (27) 0/11 (0) 0/13 (0)

124/138 (90) 13/138 (9) 1/138 (1) 4/147 (3)

29/32 (91) 0/32 (0) 3/32 (10)

1/13 (8) 1/13 (8) 11/13 (84)

2/47 (4) 4/47 (9) 41/47 (87)

29/31 (94) 34/37 (92) 3/37 (8)

13/13 (100) 1/12 (9) 11/12 (91)

146/151 (97) 5/151 (3) 146/151 (97)

6/32 (19) 4/32 (10) 22/32 (71)

9/10 (90) 0/10 (0) 1/10 (10)

105/108 (97) 3/108 (3) 0/108 (0)

28/36 (78) 8/36 (22)

11/13 (85) 2/13 (15)

131/137 (95) 6/137 (4)

25/36 (69) 11/36 (31)

9/11 (82) 2/11 (18)

101/109 (92) 8/109 (7)

10/31 (32) 21/31 (68) 0/31 (0)

4/10 (40) 6/10 (60) 0/10 (0)

124/137 (91) 13/137 (9) 0/137 (0)

a

S, susceptible; I, intermediate; R, resistant. There were 43 new taxon isolates, 24 N. brasiliensis sensu stricto isolates associated with extracutaneous diseases, and 236 N. brasiliensis sensu stricto isolates associated with cutaneous diseases. b

of the new taxon required MICs of #0.5 mg/ml, compared with 3 of 60 (5%) isolates of N. brasiliensis sensu stricto. Evaluations of other antimicrobial agents revealed a number of other susceptibility differences that are unique to the new group. Thirty-four of 37 (92%) new taxon isolates were resistant to the 25-mg trimethoprim-sulfamethoxazole disk by disk diffusion. These isolates were susceptible to sulfamethoxazole by MIC determinations (Table 3), a finding that may reflect the low sulfa concentration in the disk (23.75 mg). For amoxicillinclavulanic acid, 26 of 32 (81%) isolates were resistant or intermediate by MIC determinations, 8 of 36 (22%) were resistant by MIC determinations to cefotaxime, and 11 of 36 (31%) were resistant to ceftriaxone. Although the MICs of amikacin were still within the susceptible range, those for 21 of 31 (68%) new taxon isolates were 4 to 16 mg/ml, compared with those for 19 of 147 (13%) N. brasiliensis sensu stricto isolates. The susceptibilities to imipenem and ampicillin for these two groups were comparable (data not shown), with .90% of isolates resistant. HPLC. A typical mycolic acid pattern for Nocardia spp. was evident among both N. brasiliensis sensu stricto and new taxon isolates. A total of 10 mycolic peaks was seen (labeled A through J). Members of the new taxon (29 of 30 strains) had

NEW NOCARDIA TAXON

VOL. 33, 1995

FIG. 1. Comparison of HPLC patterns of mycolic acid–p-bromophenacyl esters from isolates of the new taxon and N. brasiliensis sensu stricto isolates, with peaks labeled A through J.

two early peaks (A and B) that were not seen with any of the 20 N. brasiliensis sensu stricto isolates studied, regardless of whether they were cutaneous or extracutaneous in origin. The majority of the N. brasiliensis sensu stricto isolates tested had peaks C through J, while the remaining isolates had peaks C through G only. Among new taxon isolates, the majority had peaks C through G only. Typical examples of the N. brasiliensis sensu stricto and new taxon HPLC patterns are shown in Fig. 1. b-Lactamase isoelectric focusing patterns. The majority (75%) of N. brasiliensis sensu stricto isolates (21 of 28, including 12 from a previous study [21]) exhibited a common doubleband b-lactamase pattern, with major band pIs of 5.00 and 5.05 (previously shown to be present in 71% of the N. brasiliensis isolates tested [21]), while 25% (7 of 28) exhibited pIs of 4.73 and 4.91. Of 12 new taxon isolates, 10 (83%) had b-lactamases whose single major band pI was 5.23, which is higher than either pattern for N. brasiliensis. The remaining two new taxon isolates produced b-lactamase patterns which resembled each of the N. brasiliensis b-lactamase pI patterns. PCR-RFLP analysis. Fourteen isolates of N. brasiliensis sensu stricto, including ATCC 19296T, and 19 isolates of the

TABLE 4. Comparison of N. brasiliensis sensu stricto and new taxon RFLP patterns with 11 enzymes RFLP band size(s) (bp)a Restriction endonuclease

BstEII AciI CfoI (HhaI) MspI HaeIII BsaHI BanII BstNI HinfI NaeI SstII

N. brasiliensis sensu stricto

New taxon

440 (not cut) 125, 110 230 185, 145, 115 210, 160 185 325 305, 145 295, 120 255, 175 305, 130

435 (not cut) 125, 110 230 180, 130 160 380 (63%), 305 (37%) 290, 105 440 (not cut) 250, 185 310, 130 445 (not cut)

a Based on calculations from multiple gels (two to six), with 2 to 7 per gel and 2 to 21 per every two to six gels. Standard deviations were for values from same gel #2 bp and #3 bp for values from multiple gels.

1531

FIG. 2. Results of PCR amplification of a 439-bp fragment of the 65-kDa heat shock protein gene, digestion with HaeIII, and agarose gel electrophoresis of N. brasiliensis sensu stricto (lanes 1 through 6) and new taxon (lanes 7 through 12) isolates. Linear base pair standards (STD) were included in the indicated lanes.

new taxon, including one isolate (N975) that was ciprofloxacin resistant and the two ATCC strains (ATCC 51511 and ATCC 51512), were restricted with 11 restriction enzymes (Table 4). Neither group was cut with BstEII, while almost identical RFLP patterns were seen with AciI and CfoI. The eight remaining restriction endonucleases gave different RFLP patterns for these two taxa. With HaeIII, for example, the N. brasiliensis sensu stricto isolates had two large fragments (210 and 160 bp), while the new taxon isolates had only one large fragment (160 bp) and three smaller ones (less than 100 bp) (Fig. 2). DISCUSSION N. brasiliensis is best known as a cause of primary cutaneous infections in both adults and children (2–4, 9, 16). In largescale reports of patients, however, approximately 10 to 30% of clinical isolates of N. brasiliensis have been associated with extracutaneous (invasive) disease (2, 4, 16), usually involving lungs. Disseminated disease has also been described (1, 2, 8, 17). While the majority of patients with primary cutaneous infections due to N. brasiliensis are immunocompetent, most extracutaneous disease has occurred in severely immunocompromised patients (1, 2, 4, 16). In a survey of infectious disease specialists from 1972 to 1974, Beaman and colleagues found that N. brasiliensis was responsible for 5.6% (11 of 197 cases) of pulmonary and systemic nocardiosis (3). In a subsequent review, Beaman and Beaman found N. brasiliensis to be responsible for 4 to 7% of all reported cases of pulmonary, CNS, and disseminated nocardiosis (2). In this series of 304 isolates of N. brasiliensis (which is the largest reported to date), the incidence of invasive disease (12% in Queensland and 22% in the United States) was similar to those in these previous reports. The relative importance of N. brasiliensis (both N. brasiliensis sensu stricto and the new taxon) among all cases of pulmonary, CNS, and disseminated nocardiosis was not determined. As with previous cases of invasive disease due to N. brasiliensis, most of the patients (98%) were immunosuppressed, most commonly as a result of corticosteroid therapy (74%). Our study found that the majority of these extracutaneous isolates (32 of 54 [59%] studied in the United States and 5 of 8 [63%] studied in Queensland) belong to a new taxon that is

1532

WALLACE ET AL.

J. CLIN. MICROBIOL.

TABLE 5. Selected tests that are useful for separating isolates of the new taxon from isolates of N. brasiliensis sensu stricto New taxon Parametera

Antimicrobial susceptibility Ciprofloxacin MIC (#0.5 mg/ml) Zone ($35 mm) Minocycline, zone (#15 mm) Clarithromycin, MIC (#0.5 mg/ml) Adenine hydrolysis Nitrate reduction PCR RFLP pattern ofc: New taxon N. brasiliensis HPLC mycolic peaks (A and B) a b c

Agreementb for all isolates

1 1 1 1 1 2

(38/40 [95]) (35/37 [94]) (31/37 [84]) (29/32 [90]) (28/33 [85]) (5/28 [19])

1 (19/19 [100]) 2 (0/19 [0]) 1 (29/30 [97])

Result for ATCC 51511

Result for ATCC 51512

0.25 mg/ml 46 mm 11 mm 0.5 mg/ml 1 2

0.25 mg/ml 44 mm 12 mm 0.063 mg/ml 1 2

1 2 1

1 2 1

Agreementb for N. brasiliensis sensu stricto

2 2 2 2 2 1

(0/123 [0]) (0/164 [0]) (4/147 [3]) (2/47 [4]) (2/47 [4]) (15/18 [83])

2 (0/14 [0]) 1 (14/14 [100]) 2 (0/20 [0])

Values in parentheses are standards for agreement. Agreement is represented as 1 (.80%) and 2 (,20%). Numbers in parentheses are number positive/number tested (percent positive). Restriction enzymes that are useful for differentiation include HaeIII, HhaIII, etc.

distinct from that of isolates of N. brasiliensis which cause cutaneous disease. On the basis of previous numbers for the incidence of N. brasiliensis (2), it is likely that isolates of the new taxon are responsible for 3 to 5% of the cases of pulmonary, CNS, and disseminated nocardiosis. Isolates of the new taxon were strongly associated with invasive disease. These 43 isolates had been recovered from lungs (70%), the CNS (23%), and/or cases of disseminated disease (37%). These percentages are similar to those reported for isolates of N. asteroides complex (2) but are highly dissimilar (as previously noted) to those for N. brasiliensis isolates (2, 4, 16). Previous studies have shown that all current Nocardia spp. have mycolic acids and that their HPLC patterns are distinct from those of other mycolic acid-containing genera, including Mycobacterium, Rhodococcus, and Corynebacterium (6). Thus far, however, individual Nocardia species have not given sufficiently unique mycolic acid patterns to allow for species identification. This may be due, in part, to the recognition that N. asteroides complex (20) actually consists of two additional recognized species (N. nova and N. farcinica) (20, 25) and probably several others (20, 24, 25). Isolates of the new taxon have an HPLC pattern that is different from that of isolates of N. brasiliensis sensu stricto in its two early peaks. More correlation between HPLC mycolic acid-ester patterns and carefully identified Nocardia spp. needs to be done to confirm the unique character of the new taxon’s pattern. PCR-RFLP analysis has recently been described as a means of species identification for both Mycobacterium (15, 19) and Nocardia spp. (11). This method utilizes PCR to amplify selected gene fragments, with subsequent restriction endonuclease digestion. Most of these studies have used the 65-kDa heat shock protein gene (11, 15, 19). We used the primers and 439-bp sequence recommended by Telenti et al. (19) for separating Mycobacterium spp. By using the same gene fragment isolates, the new taxon isolates and N. brasiliensis sensu stricto isolates gave strikingly different RFLP patterns (Fig. 2). The RFLP pattern of new taxon isolates is also unique among those of all recognized species of the genus Nocardia (16a), strongly suggesting that this new taxon is a new species. To facilitate future genetic comparisons, two strains of this new group (N80 and N224) have been deposited with the ATCC and are listed as ATCC 51511 and ATCC 51512, respectively. Clinical recognition of this new taxon is difficult because of

its growth and biochemical similarity or identity to N. brasiliensis sensu stricto. Similar problems have hampered recognition of individual species within N. asteroides complex, including N. nova and N. farcinica (20, 25). Several tests are potentially useful for recognition of new taxon isolates (Table 5). Adenine hydrolysis, which appears to be characteristic of isolates of the new taxon, is rare among Nocardia species (10, 13) and for this reason is not routinely tested. We recommend that all extracutaneous or disseminated isolates of apparent N. brasiliensis should be screened for susceptibility to ciprofloxacin as defined by disk zone of inhibition ($35 mm) or MIC (#0.5 mg/ml). Then susceptible isolates should be forwarded to reference laboratories for additional testing, such as HPLC and/or adenine hydrolysis. ACKNOWLEDGMENTS We acknowledge the assistance of Glenn Roberts in submitting isolates for study, Elizabeth Young for help in the Queensland laboratory, and Joanne Woodring for typing the manuscript. REFERENCES 1. Arduino, R. C., P. C. Johnson, and A. G. Miranda. 1993. Nocardiosis in renal transplant recipients undergoing immunosuppression with cyclosporine. Clin. Infect. Dis. 16:505–512. 2. Beaman, B. L., and L. Beaman. 1994. Nocardia species: host-parasite relationship. Clin. Microbiol. Rev. 7:213–264. 3. Beaman, B. L., J. Burnside, B. Edwards, and W. Causey. 1976. Nocardia infections in the United States, 1972–1974. J. Infect. Dis. 134:286–289. 4. Berd, D. 1973. Nocardia brasiliensis infection in the United States: a report of nine cases and a review of the literature. Am. J. Clin. Pathol. 59:254–258. 5. Butler, W. R., K. C. Jost, Jr., and J. O. Kilburn. 1991. Identification of mycobacteria by high-performance liquid chromatography. J. Clin. Microbiol. 29:2468–2472. 6. Butler, W. R., J. O. Kilburn, and G. P. Kubica. 1987. High-performance liquid chromatography analysis of mycolic acids as an aid in laboratory identification of Rhodococcus and Nocardia species. J. Clin. Microbiol. 25: 2126–2131. 7. Jost, K. C., Jr., and D. Dunbar. 1992. Automated identification of mycobacteria by high performance liquid chromatography using computer-aided pattern recognition algorithms, abstr. U-69, p. 177. In Abstracts of the 92nd General Meeting of the American Society for Microbiology 1992. American Society for Microbiology, Washington, D.C. 8. Koll, B., A. E. Brown, T. E. Kiehn, and D. Armstrong. 1992. Disseminated Nocardia brasiliensis infection with septic arthritis. Clin. Infect. Dis. 15:469– 472. 9. Lampe, R. M., C. J. Baker, E. J. Septimus, and R. J. Wallace, Jr. 1981. Cervicofacial nocardiosis in children. J. Pediatr. 99:593–595. 10. Lechevalier, H. A. 1989. Nocardioforms, p. 2348–2361. In P. H. A. Sneath, N. S. Mair, M. E. Sharpe, and John G. Holt (ed.), Bergey’s manual of

VOL. 33, 1995 systematic bacteriology, vol. 2. Williams & Wilkins, Baltimore. 11. Lungu, O., P. Della Latta, I. Weitzman, and S. Silverstein. 1994. Differentiation of Nocardia from rapidly growing Mycobacterium species by PCRRLFP analysis. Diagn. Microbiol. Infect. Dis. 18:13–18. 12. McNeil, M. M., and J. Brown. 1994. The medically important aerobic actinomycetes: epidemiology and microbiology. Clin. Microbiol. Rev. 7:357–417. 13. Mishra, S. K., R. E. Gordon, and D. A. Barnett. 1980. Identification of nocardiae and streptomycetes of medical importance. J. Clin. Microbiol. 11:728–736. 14. National Committee for Clinical Laboratory Standards. 1990. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically, 2nd ed. Approved standard M7-A2. National Committee for Clinical Laboratory Standards, Villanova, Pa. 15. Plikaytis, B. B., B. D. Plikaytis, M. A. Yakrus, W. R. Butler, C. L. Woodley, V. A. Silcox, and T. M. Shinnick. 1992. Differentiation of slowly growing Mycobacterium species, including Mycobacterium tuberculosis, by gene amplification and restriction fragment length polymorphism analysis. J. Clin. Microbiol. 30:1815–1822. 16. Smego, R. A., Jr., and H. A. Gallis. 1984. The clinical spectrum of Nocardia brasiliensis infection in the United States. Rev. Infect. Dis. 6:164–180. 16a.Steingrube, V. A., et al. Unpublished data. 17. Steingrube, V. A., R. J. Wallace, Jr., B. A. Brown, Y. Pang, B. Zeluff, L. C. Steele, and Y. Zhang. 1991. Acquired resistance of Nocardia brasiliensis to clavulanic acid related to a change in b-lactamase following therapy with amoxicillin-clavulanic acid. Antimicrob. Agents Chemother. 35:524–528. 18. Steingrube, V. A., R. J. Wallace, Jr., B. A. Brown, Y. Zhang, L. C. Steele, G. Young, and D. R. Nash. 1993. Partial characterization of Nocardia farcinica b-lactamases. Antimicrob. Agents Chemother. 37:1850–1855.

NEW NOCARDIA TAXON

1533

19. Telenti, A., F. Marchesi, M. Balz, F. Bally, E. Bo¨ttger, and T. Bodmer. 1993. Rapid identification of mycobacteria to the species level by polymerase chain reaction and restriction enzyme analysis. J. Clin. Microbiol. 31:175–178. 20. Wallace, R. J., Jr., B. A. Brown, M. Tsukamura, J. M. Brown, and G. O. Onyi. 1991. Clinical and laboratory features of Nocardia nova. J. Clin. Microbiol. 29:2407–2411. 21. Wallace, R. J., Jr., D. R. Nash, W. K. Johnson, L. C. Steele, and V. A. Steingrube. 1987. b-Lactam resistance in Nocardia brasiliensis is mediated by b-lactamase and reversed in the presence of clavulanic acid. J. Infect. Dis. 156:959–966. 22. Wallace, R. J., Jr., E. J. Septimus, T. W. Williams, Jr., R. H. Conklin, T. K. Satterwhite, M. B. Bushby, and D. C. Hollowell. 1982. Use of trimethoprimsulfamethoxazole for treatment of infections due to Nocardia. Rev. Infect. Dis. 4:315–325. 23. Wallace, R. J., Jr., and L. C. Steele. 1988. Susceptibility testing of Nocardia species for the clinical laboratory. Diagn. Microbiol. Infect. Dis. 9:155–166. 24. Wallace, R. J., Jr., L. C. Steele, G. Sumter, and J. M. Smith. 1988. Antimicrobial susceptibility patterns of Nocardia asteroides. Antimicrob. Agents Chemother. 32:1776–1779. 25. Wallace, R. J., Jr., M. Tsukamura, B. A. Brown, J. Brown, V. A. Steingrube, Y. Zhang, and D. R. Nash. 1990. Cefotaxime-resistant Nocardia asteroides strains are isolates of the controversial species Nocardia farcinica. J. Clin. Microbiol. 28:2726–2732. 26. Wallace, R. J., Jr., K. Wiss, R. Curvey, P. H. Vance, and J. Steadham. 1983. Differences among Nocardia spp. in susceptibility to aminoglycosides and b-lactam antibiotics and their potential use in taxonomy. Antimicrob. Agents Chemother. 23:19–21.