- Email: [email protected]
Identification of Penicillin and other Beta-Lactam Resistance in Streptococcus Pneumoniae by Polymerase Chain Reaction
Identification of Penicillin and other Beta-Lactam Resistance in Streptococcuspneumoniaeby Polymerase Chain Reaction K i m i k o U b u k a t a , * T o m o k o M u r a k i , A t s u m i Igarashi, Yasuko Asahi, M a s a t o s h i K o n n o Department of Clinical Pathology, Teikyo Uni',ersity School of Medicine, Tokyo, Japan To identify penicillin (Pc) and other ~lactam resistance in 310 clinical isolates of Streptococcuspneumoniae by polymerase chain reaction (PCR), 3 sets of primers were designed to amplify Pc-binding protein (PBP) genes previously detected in Pc-susceptible strains: 1) a 430-bp fragment of the pbpla gene, 2) a 292-bp fragment of the pbp2x gene, and 3) a 77-bp fragment of the pbp2b gene. The amplified regions of each PBP gene were positioned in highly divergent sequences of Pc-resistant S. pneumoniae. In other words, isolates for which these DNA fragments were detected were regarded as possessing sequences almost the same as that of the susceptible R6 strain and those for which these DNA fragments were not detected were assumed to have mutations. A set of primers that amplify 273 bp of the autolysin (lytA) gene to identify S. pneumoniae was applied as well. Of 166 isolates for which the minimum inhibitory concentration (MIC) of Pc were < 0.06 pg/mL, 83 (50.0%) were confirmed to be true susceptible strains with no PBP gene mutation and most of the remaining strains were found to possess pbp2x mutation. In contrast, most of 109 isolates for which the MIC of Pc were > 0.5 #g/mL were confirmed to possess mutations in all three PBP genes. Thirty-five strains for which the MIC of Pc ranged from 0.125 to 0.25 pg/mL possessed various PBP gene mutations. The relationships between susceptibilities to 9 fl-lactams of S. pneumoniae and PBP gene mutation were analyzed by multiple regression analysis. Antibiotics were classified into 4 types according to the differences in PBP gene mutation affecting their MIC levels, 1) the MIC of Pc and ampicillin were affected by pbpla and pbp2b mutations; 2) those of cefotaxime, cefpodoxime, and cefditoren were affected clearly by pbp2x mutation; 3) those of cefaclor and cefdinir were affected more strongly by pbpla mutation than the pbp2x; and 4) the MIC of faropenem and imipenem were affected strongly by pbp2b mutation.These findings suggest that it may be possible to easily determine whether a S. pneumoniae isolate is susceptible or resistant to Pc, cefotaxime, and other ~lactams by applying PCR using a combination of primers. J Infect Chemother 1997;3:190-197 Key words: Streptococcuspneumoniae, PCR, penicillin resistance, PBP genes, fl-lactams
INTRODUCTION Severe infections due to penicillin (Pc) and cefotaxime resistant Streptococcus pneumoniae have become a serious problem worldwide. It requires at least 3 days to obtain the results of an antibiotic susceptibility test for S. pneumoniae isolates. In addition, the results of M I C m e a s u r e m e n t by biological m e t h o d s , such as the microdilution and agar plate dilution methods, always vary. In our previous studies, therefore, we attempted to use P C R for the rapid identification of Pc-resistant S. pneumoniae (PRSP) from clinical specimens. A m o n g the pbpla, pbp2x, and pbp2b genes involved in Pc resistance, mutations of the pbp2b gene were detected by PCR. As a result, we found that 70.3% of S. pneumoniae having intermediate Pc resistance (PISP) and PRSP for which the M I C of Pc was ___0.125 p g / m L had a class B mutation in the pbp2b gene while only 1.8% of PISP and PRSP had a class A mutation. The remaining 27.9% of
the isolates could not be definitely identified using PCR. 1 Some of these isolates were later found to have a characteristic 9-nucleotide insertion in the pbp2b gene. 2 Therefore, to improve the probability of P I S P / P R S P detection, the pbpla and pbp2x genes were examined along with the pbp2b gene. T h e sequences of pbpla, 3 pbp2b, 2,4-6 and pbp2x 7 genes of PRSP were more highly diverged than those of Pc-susceptible S. pneumoniae (PSSP). Accordingly, we designed the primers to amplify a part of normal PBP genes in PSSP where several mutations were c o m m o n l y found in resistant strains. These positions were near the conserved amino acid sequences which are essential for the transpeptidase activity of PBP. Specifically, we attempted to detect only true fl-lactam-susceptible isolates. In this paper we describe the relationship, revealed by multivariate anylysis, between the results ofpbpla, pbp2x, and pbp2b gene amplifications by P C R and the susceptibility to fl-lactams of clinical isolates of S. pneumoniae.
MATERIALS A N D METHODS
ReceivedJul. 8, 1997; accepted for publication in revised form Aug. 27, 1997.*Correspondence and requests for reprints to: Department of Clinical Pathology, Teikyo University School of Medicine, 2-I1-1 Kaga, Itabashi-ku, Tokyo 173, Japan. 190
Strains Clinical isolates (n = 310) were obtained from 49 institutions participating in the Working G r o u p for PRSP.
1341-321X/97/0304-0190/US$3.009 JSC/CLJ 1997
Rapid detection of Beta-lactam resistance in S. pneumoniae
These isolates were collected from various specimens including cerebrospinal fluid, blood, pleural fluid, pharynx, nasopharynx, sputum, and otorrhea between October 1995 and March 1997. Isolates were grown on sheep blood agar (Nippon Becton Dickinson, Tokyo, Japan) at 37~ with 5% CO 2 and a single colony was purified. They were stored in 10% skim milk (Difco Laboratories, Detroit, MI) at 80~ Identification of isolates as S. pneumoniae was confirmed by PCR of the autolysin (lytA) gene in addition to optochin, bile solubility, and inulin fermentation.
Susceptibility tests Using the agar plate dilution method, susceptibilities of S. pneumoniae to ~-lactams were determined using cation-adjusted Muetler Hinton agar (Eiken, Tokyo, Japan) supplemented with 5 % defibrinized sheep blood. Bacteria inocula were prepared following a previously reported method.1 ~-lactams employed in this study were Pc, ampicillin (Amp), and cefditoren (Cdtr) (Meiji Seika, Tokyo, Japan), Ctx (Nippon Roussel, Tokyo, Japan), cefaclor (Cch Shionogi, Osaka, Japan), cefdinir (Cfdn: Fujisawa Pharmaceutical, Osaka, Japan), cefpodoxime (Cpdx: Sankyo, Tokyo, Japan), faropenem (Frpm: Suntory, Tokyo, Japan), and imipenem (Ipm: Banyu Pharmaceutical, Tokyo, Japan).
PCR primers The sequences of 4 sets of primers used for PCR are shown in Table 1. To confirm that isolates were S. pneumoniae, the lytA gene 8 that encodes autolysin was simultaneously amplified. The oligonucleotide primers for detection of three PBP genes were designed to amplify parts of the pbpla, 3pbp2x,9 and pbp2b genes 2,1~only in susceptible strains.These parts were positioned in blocks of highly diverged sequences previously identified in the mosaic PBP genes of Pc non-susceptible S. pneumonfig.e. 2-4'9
Primer mixture A contained the primers for detecting the lytA and pbpla genes, and primer mixture B contained primers for detecting the pbp2x and pbp2b genes.
PCR conditions A single colony of S. pneumoniae grown on a blood agar plate was suspended in a 0.5 m L microtube containing 30 pL of lysis solution. The composition of the lysis solution was previously reported. 1The tubes were put in a thermal cycler (Gene Amp PCR System 9600-R, PerkinElmer, Norwalk, CT, USA), and the bacterial cells were lysed for 10 minutes at 60~ and for 5 minutes at 94~ Next, 1 p L of the bacterial lysate was added to each of two tubes marked A (containing primer mixture A) and B (containing primer mixture B). Each tube contained 25 ~tL of reaction mixture which consisted of 60 ng of the appropriate primer, 8 pL of 25 m M of each d N T P mixture, 2.5 U of Taq DNA polymerase (Takara Biomedicals, Kyoto, Japan), and 10 pL of 10 • PCR buffer (pH 8.3) per 100 pL solution. The PCR cycling conditions were 30 cycles at 94~ for 20 seconds, 57~ for 20 seconds, and 72~ for 15 seconds. Amplified DNA fragments were analyzed using gel electrophoresis with 3% agarose.
Detection of PBP To confirm that the gene mutations were accurately detected, the PBPs of S. pneumoniae were labeled with [3H]benzylpenicillin (specific activity, 407 Gbq/mmol; Amersham, Buckinghamshire, England) and subjected to SDS-polyacrylamide gel electrophoresis (PAGE) according to the previously described method. 1 Acrylamide solutions with the following compositions were used in SDS-PAGE: the running gel modified to 10% (wt/vol) acrylamide, 0.063% (wt/vol) N, N ' methylenebisacrylamide, 0.375 MTris-HC1 buffer (pH 8.8), 0.1% SDS, 0.1% T E M E D , and 0.1% APS (wt/
Table 1. Sequences of oligonucleotide primers. Name (gene) Autolysin (/ytA)
Sequence (5' to 3')
Position
ALYI
TGAAGCGGATTATCACTGGC
694-713
ALY2
GCTAAAC TCC CTGTATCAAGCG
966-945
Product length (bp)
273
PBP 1A (pbpla) PBPI AI
AAACAAGGTCGGACTCAACC
2256-2275
PBPI A2
AGGTGC TACA/hATTGAGAGG
2685-2666
PB P2 X I
CCAGGTTCCACTATGAAAGTG
1003-1023
PBP2X2
CATCCGTCAAACCGAAACGG
1294-I 275
PBP281
CAATCTAGAGTCTGCTATGGA
1636-I 656
PBP2B2
GGTCAATTCCTGTCGCAGTA
1712-I 693
430
PBP 2X (pbpBx) 292
PBP 2B (pbp2b) 77
Based on data in reference 8 (lytA), 3 (pbpla), 9 (pbp2x), and 10 (pbp2b).
191
J Infect Chemother 1997;3:190 197
vol); and the stacking gel modified to 4% (wt/vol) acrylamide, 0.063% N, N'-methylenebisacrylamide, 0.125 MTris-HC1 buffer (pH 6.8), 0.1% SDS, 0.1% T E M E D , and 0.1% APS (all reagents: Wako Pure Chemical, Tokyo, Japan). T h e voltages for SDS-PAGE were 60 V in the stacking gel and 90 V in the running gel.
Statistical analysis Statistical analysis was performed using Microsoft Excel Multivariate Analysis software (version 3.0). The partial regression coefficients, fixed values, and standard partial regression coefficients (Table 2) were obtained with multiple regression analysis.
RESULTS
Amplified DNA profile Figure 1 shows PCR-amplified DNA profiles obtained from 9 strains of S. pneumoniae.Three DNA bands, 430 bp in column A, and 292 bp and 77 bp in column B, correspond to the DNA fragments amplified with the primers that were designed from the sequences of the pbpla, pbp2x, and pbp2b genes of PSSP R6 strain. The 273 bp DNA fragment detected in column A correspond to the products of lytA gene for identification of S. pneumoniae species. When each DNA band ofpbpla, pbp2x, and pbp2b was detected, these PBP genes were regarded as having al-
Table 2. M u l t i p l e regression analysis between the presence of mutations of PBP genes and MICs of 9 fl-lactam antibiotics. Penicillin G
Ampicillin
Cefac]or Cefotaxime Cefpodoxime Cefdinir Cefditoren
Faropenem Imipenem
2.3135 1.0363 2.3359 -5.8835
2.1174 1.1839 2.3145 -5.2644
2.9277 1.1654 1.8175 -0.7587
1.4086 2.6813 0.5534 -5.9134
1.6985 2.9110 0.7959 -4,6045
2.0604 2.0003 1.0924 -3.8528
1.4863 2.7064 0.3658 -6.1469
1.6173 0.5339 2.2680 -6.5341
1.6480 0.5472 2.0861 -7.4817
0.4437 0.1841 0.4450
0.4144 0.2146 0.4500
0.5243 0.1933 0.3233
0.3348 0.5903 0.1307
0.3514 0.5578 0.1636
0.4283 0.3851 0.2255
0.3660 0.6174 0,0895
0.3784 0.1157 0.5270
0.3987 0.1226 0,5013
0.6066 0,4456 0.6231
0.5670 0.4889 0.6141
0.5907 0.3907 0.4253
0.3859 0.7573 0.1675
0.4163 0.7525 0.2166
0.4446 0.5744 0.2626
0.4520 0.8023 0.1279
0.4789 0.2537 0.6204
0.4950 0.2656 0.5977
0.8938 0.5767 0.8806
0.8831 0.5952 0.8741
0.8797 0.5751 0.8282
0.7234 0.8034 0.6423
0.7512 0.7922 0.6756
0.7959 0.6813 0.7301
0.7341 0.8287 0.6375
0.8625 0.5103 0.8818
0.8653 0.5165 0.8755
Coefficient of determination
0.8946
0.8870
0.8401
0.8003
0.8146
0.7679
0.8374
0.8501
0.8473
Multiple correlation co.*
0.8936
0.8859
0.8386
0.7984
0.8146
0.7657
0.8358
0.8486
0.8458
Partial regression coefficient
pbpla pbp2x pbp2b Fixed value Standard partial reg. coefficient
pbpla pbp2x pbp2b PartiaI correlation coefficient
pbp la pbp2x pbp2b Correlation coefficient
pbpla pbp2x pbp2b
* Multiple correlation coefficient adjusted for the degree of freedom.
Fig.1 Agarose gel electrophoresis of PCR-amplified D N A fragments of the lytA (273 bp) and pbpla (430-bp) genes in column A and pbp2x (292-bp) and pbp2b (77-bp) genes in column B in S. pneumoniae. These 3 PBP genes with the same sequences as penicillin-susceptible R6 strain 3,9,1~were amplified.
192
Fig.2 Fluorogram of the PBP of S. pneumoniae. M e m b r a n e fractions are solubilized after a 10 minute incubation with [3H]benzylpenicillin. MIC of benzylpenicillin and the results of PCR amplification (+: positive; - : negative) are listed b e l o w the illustration.
Rapid detection of Beta-lactam rosistance in S. pneumoniae
most the same sequences as those of the R6 strain of PSSE In contrast, when the D N A bands were not detected or different sizes were detected, their genes were regarded as possessing sequences different from those of the R6 strain.
Correctness of PCR detection of PBP genes To determine whether the presence or absence of mutations in the PBP genes had been correctly detected by PCR, the affinities of PBPs from the 25 strains selected at random were examined with [3H]benzylpeniciUin. The results for the 9 strains are shown in Fig. 2 together with the M I C of Pc and the results of P C R . W h e n the mutation(s) in the pbpla gene were detected, the rate of flow (Rf) value of PBP 1A from the mutationalpbpla was a little larger than that of the normal gene, and/or its affinity for [3H]benzylpenicitlin was decreased. PBP 2X from the mutatedpbp2x had a larger Rfvalue than did PBP 2A in this conditions of SDS-PAGE. When the mutation(s) was present in the pbp2b gene, PBP 2B was scarcely detected or its affinity for [3H]benzylpenicillin seemed to be reduced. From these findings, we concluded that the P C R using these sets of primers correctly reflects the changes of affinity in each PBE
Penicillin susceptibility and PBP genes Figure 3 shows the distribution of the M I C of Pc and the P C R results of pbpla, pbp2x, and pbp2b genes by P C R for the 310 strains of S. pneumoniae.
O f the 166 PSSP strains for which the M I C were _<0.063 pg/mL, 83 (50.0%) possessed normal genes of pbpla, pbp2x, and pbp2b. A m o n g the remaining 83 strains, 72 (43.4%) had a mutation in thepbp2x gene, and 1 (0.6%) in thepbp2b gene, 8 (4.8%) in thepbpla andpbp2x genes, and 2 (1.2%) in the pbp2x and pbp2b genes. O n the contrary, 104 (95.4%) o f the 109 P R S P strains for which the M I C was >_0.5 flg/mL had mutations in the three genes, pbpla, pbp2x, and pbp2b. The remaining 5 (4.6%) strains had a mutation in the pbpla and pbp2x gene. T h e 35 PISP strains for which the M I C ranged from 0.125 to 0.25 p g / m L possessed mutations in the pbpla, pbp2x, and pbp2b genes with various combinations, such as pbpla and pbp2b (11 strains), pbpla and pbp2x (4 strains), pbp2x and pbp2b (12 strains), etc.
Relationship between PBP ggnes and fl-lactam susceptibility Figure 4 shows the relationship between fl-lactam susceptibility and the pbpla, pbp2x, and pbp2b genes for the 310 pneumococcal strains. Two types were observed in these 8 fl-lactams. One is the Pc type by which the strains with pbp2x mutation were not separated from susceptible strains; ampicillin, c e f a c l o r , f a r o p e n e m , and imipenem belonged to this Pc type. A n o t h e r is the cefotaxime type by which the strains with pbp2x mutation were separated from susceptible strains with the
Fig.3 Correlation between MIC of benzylpenicillin and PBP gene mutations for a total 310 isolates of S. pneumoniae.
193
J Infect Chemother 1997;3:190- 797
Fig.4 Correlation between MIC of 8 ~-Iactam antibiotics and PBP gene mutations of 310 isolates of S. pneumoniae.
194
Rapid detection of Beta-lactam resistance in S. pneumoniae
M I C shifted to resistant side with 4 to 8 times higher values t h a n the f o r m e r ; cefdinir, c e f p o d o x i m e , and cefditoren belonged to this cefotaxime type.
Influence of PBP gene mutation on M I C level of fllactams Multiple regression analysis was applied to determine whether the presence of mutation of the 3 PBP genes influenced the M I C level of fl-lactams.Values of the M I C of fl-lactams for 310 p n e u m o c o c c a l strains changed to logarithm based 2 and the presence or absence of m u -
tation in pbpla, pbp2x, and pbp2b were used as a criterion variable and explanatory variable, respectively. Table 2 shows the partial regression coefficients, fixed values, and standard partial regression coefficients for the pbpla, pbp2x, and pbp2b obtained by the analysis. It also shows the coefficients of determination and multiple correlation coefficients adjusted for the degrees of freedom that indicate the accuracy of this analysis. Standard partial regression coefficients that indicate the weight of the explanatory variable showed high values forpbpla and pbp2b in Pc and ampicillin, for pbp2b in faropenem
Table 3. Estimated MIC value of ~ l a c t a m antibiotics calculated from multiple regression formula. Pattern of PBP gene mutation
Estimated MIC value (,u.g/mL) Cefaclor Cefotaxime Cefpodoxime Cefdinir C~fditoren
Penicillin G
Ampicillin
Normal
0.0169
0.0260
0.5910
0.0166
0.0411
0.0692
0.0141
Faropenem Imipenem 0.0141
0.0056
pbpla pbp2x pbp2b
0.0842 0.0347 0.0855
0.1129 0.0591 0.1294
4.4971 1.3257 2.0832
0.0440 0.1064 0.0243
0.1334 0.3062 0.0714
0.2887 0.2774 0.1476
0.0395 0.0921 0.0182
0.0433 0.0204 0.0680
0.0175 0.0082 0.0238
pbpla + pbp2x pbpla +pbp2b pbp2x + pbp2b
0.1727 0.4251 0.1754
0.2565 0.5616 0.2940
10.0867 5.8510 4.6725
0.2815 0.0646 0.1562
1.0035 0.2316 0.5368
1.1572 0.6156 0.5916
0.2581 0.0509 0.1187
0.0627 0.2085 0.0984
0.0256 0.0744 0.0347
pbpla+pbp2x+pbp2b
0.8719
1.2757
35.5530
0.4146
1.7422
2.4674
0.3325
0.3019
0.1088
Fig. 5 Predicted MIC values of fl-lactams calculated by inserting the variables of the ~ l a c t a m s and the PBP gene mutation into the multiple regression formula.
195
J Infect Chemother 1997;3:190-197
and imipenem, for pbp2x in cefotaxime, cefpodoxime, and cefditoren, and for pbpla in cefaclor and cefdinir. T h e value of multiple correlation coefficient adjusted for degrees of freedom ranged from 0.7657 for cefdinir to 0.8936 for Pc. Based on multiple regression analysis, the formula " M I C (log2) -- partial regression coefficient ofpbpla x mutation of pbpla + partial regression coefficient of pbp2x x mutation of pbp2x + partial regression coefficient ofpbp2b x mutation ofpbp2b - fixed value" holds good. Predicted M I C of/%lactams were calculated easily by inserting appropriate variables in the formula. Table 3 shows the M I C values estimated for 9 /3lactams that were calculated by entering the presence of mutations inpbpla, pbp2x, andpbp2b genes into the formula. These values are expected to be fit to the actual M I C value with a high probability when PBP gene mutation is demonstrated definitely by PCR. Figure 5 shows the predicted M I C of Pc, cefotaxime, cefaclor, and faropenem in cases of PBP gene mutation being observed in combinations. Nine antibiotics were classified into these 4 types. That is, the M I C of the Pc type including ampicillin were affected more strongly by the mutation ofpbpla and pbp2b than that ofpbp2x. In contrast, the M I C of the cefotaxime type including cefpodoxime and cefditoren were apparently more strongly affected by the mutation ofpbp2x than others. It was r e m a r k e d that the M I C o f f a r o p e n e m and imipenem were distinctly affected by pbp2b mutation, and those of cefaclor and cefdinir by pbpla mutation.
DISCUSSION T h e resistance ofS. pneumoniae to Pc and cefotaxime has been shown to be associated with mosaic m u t a tions in the pbpla, 11-13 pbp2b, 12-14 a n d pbp2x 7,15 genes. O f these genes, the m u t a t i o n s in the pbp2b mediate a low-level resistance to penicillin and pbp2x cefotaxime respectively. 6,16 A d d i n g pbpla m u t a t i o n s seems to cause a high resistance. 15 T h e s e findings suggest that rapid identification of P S S P and P I S P / P R S P m a y be possible by the detection of the m u t a tions in P B P genes. Pursuing this possibility, we investigated the relationship between penicillin susceptibility and the pbp2b gene. We found that 89.8% of the isolates for which the M I C was _> 1.0 p g / m L had a class B mutation of the pbp2b gene.1 F r o m these findings, it was suggested that the c o n c o m i t a n t detection of pbp2b and the lytA genes should allow us to identify simultaneously S. pneumoniae and PRSP. However, it was impossible to identify 10.2% of the isolates as PRSP, although the M I C of these isolates for penicillin was --_ 1.0 p g / m L and the affinity of PBP 2B was clearly decreased. Analysis of the sequences ofpbp2b genes in these strains showed that a 9-nucleotides repetition was inserted between the S W K and SSN of con-
196
served amino acid sequences in the transpeptidase region. 2 Many other types of variants were found. To raise the percentage of PRSP detection to compensate for this variety o f g e n e mutations, primers were designed to amplify the genes of susceptible strains but not of resistant strains. The clinical isolate was taken as PSSP when the D N A amplification occurred and as PISP or PRSP when it did not occur. The results clarified that mutations in the 3 PBP genes affected M I C . In addition, the multivariate analysis showed that the influence of PBP gene mutations on M I C differed according to the type of antibiotic. In other words, for penicillins, penem, and carbapenems, the pbpla and pbp2b gene mutations strongly affected M I C , while the pbp2x gene mutation had little influence. On the other hand, it was found that the pbp2x mutation h a d the strongest influence on the M I C o f m a n y cephems.The M I C of cefac~lor and cefdinir was strongly affected by the pbpla gene mutation. It is thus suggested that it is necessary to distinguish between penicillin resistance and cephem resistance for S. pneumoniae. This fact also means that the use of either a disc containing 1/lg ofoxacillin or a disc containing 2 units of Pc alone cannot distinguish correctly between the P C and cephem resistances. M I C for ]3-1actams can be predicted with a high probability by determining whether or not the pbpl a, pbp2x, and pbp2b gene mutations exist and. entering the results as explanatory variables in the multiple regression formula obtained by multivariate analysis. In clinical isolates assumed to be S. pneumoniae, examination of the 3 PBP genes together with lytA gene gives predicted values for susceptibility within 2 hours and is considered to be remarkably useful for the treatment of infectious diseases caused by & pneumoniae.
REFERENCES 1 Ubukata K, Asahi Y, Yamane A, Konno M. Combinational detection of autolysin and penicillin-binding protein 2B genes of Streptococcuspneumoniae by PCR. J Clin Microbiol 1996;34:592-596. 2 Yamane A, Nakano H, Asahi Y, Ubukata K, Konno M. Directly repeated insertion of 9-nucleotide sequence detected in penicillin-binding protein 2B gene of penicillin-resistant Streptococcuspneumoniae. Antimicrob Agents Chemother 1996;40:1257 1259. 3 Martin C, Sibold C, Hakenbeck R. Relatedness of penicillin-binding protein 1a genes from different clones of penicillin-resistant Streptococcus pneumoniae isolated in South Africa and Spain. EMBO J 1992; 11:3831-3836. 4 Dowson CG, Hutchison A, Brannigan JA, George RC, Hansman D, Lifiares J, et al. Horizontal transfer of penicillin-binding protein genes in penicillin-resistant clinical isolates of Streptococcus pneumoniae. Proc Natl Acad Sci USA 1989;86:8842-8846. 5 Dowson CG, Hutchison A, Spratt BG. Extensive re-modelling of the transpeptidase domain of penicillin-binding protein 2B of a penicillin-resistant South African isolate of Streptococcuspneumoniae. Mol Microbiol 1989;3:95-102.
Rapid detection of Beta-factam resistance in S. pneumoniae
6 Smith AM, Klugman KP. Alterations in penicillin-binding protein 2B from penicillin-resistant wild-type strains of Streptococcuspneumoniae. Antimicrob Agents Chemother 1995;39:859-867. 7 L a i b l e G , S p r a t t B G , H a k e n b e c k R. I n t e r s p e c i e s recombinational events during the evolution of altered PBP 2x genes in penicillin-resistant clinical isolates of Streptococcuspneumoniae. Mol Microbiol 1991 ;5:1993 2002. 8 Garcia P, Garcia JL, Garcia E, Lopez R. Nucleotide sequence and expression of the pneumococcal autolysin gene f r o m its own p r o m o t e r in Escherichia coli. G e n e 1986;43:265-272. 9 Laible G, Hakenbeck R, Sicard MA, Joris B, Ghuysen JM. Nucleotide sequences of the pbpX genes encoding the p e n i c i l l i n - b i n d i n g p r o t e i n s 2X from Streptococcus pneumoniae R6 and a cefotaxime-resistant mutant, C506. Mol Microbiol 1989;3:1337-1348. 10 Dowson CG, Hutchison A, Spratt BG. Nucleotide sequence of the penicillin-binding protein 2B gene of Streptococcus pneumoniae strain R6. N u c l e i c Acids Res 1989;17:7518. 11 Hakenbeck R, Ellerbrok H, Briese T, Handwerger S, Tomasz A. Penicillin-binding proteins of penicillin-sus-
12
13 14
15
16
ceptible and -resistant pneumococci: immunological relatedness of altered proteins and changes in peptides carrying the fl-lactam binding site. Antimicrob Agents Chemother 1986;30:553-558. Jabes D, Nachman S, Tomasz A. Penicillin-binding protein families: evidence for the clonal nature of penicillin resistance in clinical isolates ofpneumococci. J Infect Dis 1989;159:16-25. Markiewicz Z, Tomasz A. Variation in penicillin-binding protein patterns of penicillin-resistant clinical isolates of pneumococci. J Clin Microbiol 1989;27:405-410. Hakenbeck R, Tarpay M, Tomasz A. Multiple changes of penicillin-binding proteins in penicillin-resistant clinical isolates of Streptococcus pneumoniae. Antimicrob Agents Chemother 1980;17:364-371. Mufioz R, Dowson CG, Daniels M, CoffeyTJ, Martin C, Hakenbeck R, et al. Genetics of resistance to third-generation cephalosporins in clinical isolates of Streptococcus pneumoniae. Mol Microbiol 1992;6:2461-2465. GrebeT, Hakenbeck R. Penicillin-binding proteins 2b and 2x of Streptococcus pneuma~ae are primary resistance determinants for different classes of fl-lactam antibiotics. Antimicrob Agents Chemother 1996;40:829-834.
197
INTRODUCTION Severe infections due to penicillin (Pc) and cefotaxime resistant Streptococcus pneumoniae have become a serious problem worldwide. It requires at least 3 days to obtain the results of an antibiotic susceptibility test for S. pneumoniae isolates. In addition, the results of M I C m e a s u r e m e n t by biological m e t h o d s , such as the microdilution and agar plate dilution methods, always vary. In our previous studies, therefore, we attempted to use P C R for the rapid identification of Pc-resistant S. pneumoniae (PRSP) from clinical specimens. A m o n g the pbpla, pbp2x, and pbp2b genes involved in Pc resistance, mutations of the pbp2b gene were detected by PCR. As a result, we found that 70.3% of S. pneumoniae having intermediate Pc resistance (PISP) and PRSP for which the M I C of Pc was ___0.125 p g / m L had a class B mutation in the pbp2b gene while only 1.8% of PISP and PRSP had a class A mutation. The remaining 27.9% of
the isolates could not be definitely identified using PCR. 1 Some of these isolates were later found to have a characteristic 9-nucleotide insertion in the pbp2b gene. 2 Therefore, to improve the probability of P I S P / P R S P detection, the pbpla and pbp2x genes were examined along with the pbp2b gene. T h e sequences of pbpla, 3 pbp2b, 2,4-6 and pbp2x 7 genes of PRSP were more highly diverged than those of Pc-susceptible S. pneumoniae (PSSP). Accordingly, we designed the primers to amplify a part of normal PBP genes in PSSP where several mutations were c o m m o n l y found in resistant strains. These positions were near the conserved amino acid sequences which are essential for the transpeptidase activity of PBP. Specifically, we attempted to detect only true fl-lactam-susceptible isolates. In this paper we describe the relationship, revealed by multivariate anylysis, between the results ofpbpla, pbp2x, and pbp2b gene amplifications by P C R and the susceptibility to fl-lactams of clinical isolates of S. pneumoniae.
MATERIALS A N D METHODS
ReceivedJul. 8, 1997; accepted for publication in revised form Aug. 27, 1997.*Correspondence and requests for reprints to: Department of Clinical Pathology, Teikyo University School of Medicine, 2-I1-1 Kaga, Itabashi-ku, Tokyo 173, Japan. 190
Strains Clinical isolates (n = 310) were obtained from 49 institutions participating in the Working G r o u p for PRSP.
1341-321X/97/0304-0190/US$3.009 JSC/CLJ 1997
Rapid detection of Beta-lactam resistance in S. pneumoniae
These isolates were collected from various specimens including cerebrospinal fluid, blood, pleural fluid, pharynx, nasopharynx, sputum, and otorrhea between October 1995 and March 1997. Isolates were grown on sheep blood agar (Nippon Becton Dickinson, Tokyo, Japan) at 37~ with 5% CO 2 and a single colony was purified. They were stored in 10% skim milk (Difco Laboratories, Detroit, MI) at 80~ Identification of isolates as S. pneumoniae was confirmed by PCR of the autolysin (lytA) gene in addition to optochin, bile solubility, and inulin fermentation.
Susceptibility tests Using the agar plate dilution method, susceptibilities of S. pneumoniae to ~-lactams were determined using cation-adjusted Muetler Hinton agar (Eiken, Tokyo, Japan) supplemented with 5 % defibrinized sheep blood. Bacteria inocula were prepared following a previously reported method.1 ~-lactams employed in this study were Pc, ampicillin (Amp), and cefditoren (Cdtr) (Meiji Seika, Tokyo, Japan), Ctx (Nippon Roussel, Tokyo, Japan), cefaclor (Cch Shionogi, Osaka, Japan), cefdinir (Cfdn: Fujisawa Pharmaceutical, Osaka, Japan), cefpodoxime (Cpdx: Sankyo, Tokyo, Japan), faropenem (Frpm: Suntory, Tokyo, Japan), and imipenem (Ipm: Banyu Pharmaceutical, Tokyo, Japan).
PCR primers The sequences of 4 sets of primers used for PCR are shown in Table 1. To confirm that isolates were S. pneumoniae, the lytA gene 8 that encodes autolysin was simultaneously amplified. The oligonucleotide primers for detection of three PBP genes were designed to amplify parts of the pbpla, 3pbp2x,9 and pbp2b genes 2,1~only in susceptible strains.These parts were positioned in blocks of highly diverged sequences previously identified in the mosaic PBP genes of Pc non-susceptible S. pneumonfig.e. 2-4'9
Primer mixture A contained the primers for detecting the lytA and pbpla genes, and primer mixture B contained primers for detecting the pbp2x and pbp2b genes.
PCR conditions A single colony of S. pneumoniae grown on a blood agar plate was suspended in a 0.5 m L microtube containing 30 pL of lysis solution. The composition of the lysis solution was previously reported. 1The tubes were put in a thermal cycler (Gene Amp PCR System 9600-R, PerkinElmer, Norwalk, CT, USA), and the bacterial cells were lysed for 10 minutes at 60~ and for 5 minutes at 94~ Next, 1 p L of the bacterial lysate was added to each of two tubes marked A (containing primer mixture A) and B (containing primer mixture B). Each tube contained 25 ~tL of reaction mixture which consisted of 60 ng of the appropriate primer, 8 pL of 25 m M of each d N T P mixture, 2.5 U of Taq DNA polymerase (Takara Biomedicals, Kyoto, Japan), and 10 pL of 10 • PCR buffer (pH 8.3) per 100 pL solution. The PCR cycling conditions were 30 cycles at 94~ for 20 seconds, 57~ for 20 seconds, and 72~ for 15 seconds. Amplified DNA fragments were analyzed using gel electrophoresis with 3% agarose.
Detection of PBP To confirm that the gene mutations were accurately detected, the PBPs of S. pneumoniae were labeled with [3H]benzylpenicillin (specific activity, 407 Gbq/mmol; Amersham, Buckinghamshire, England) and subjected to SDS-polyacrylamide gel electrophoresis (PAGE) according to the previously described method. 1 Acrylamide solutions with the following compositions were used in SDS-PAGE: the running gel modified to 10% (wt/vol) acrylamide, 0.063% (wt/vol) N, N ' methylenebisacrylamide, 0.375 MTris-HC1 buffer (pH 8.8), 0.1% SDS, 0.1% T E M E D , and 0.1% APS (wt/
Table 1. Sequences of oligonucleotide primers. Name (gene) Autolysin (/ytA)
Sequence (5' to 3')
Position
ALYI
TGAAGCGGATTATCACTGGC
694-713
ALY2
GCTAAAC TCC CTGTATCAAGCG
966-945
Product length (bp)
273
PBP 1A (pbpla) PBPI AI
AAACAAGGTCGGACTCAACC
2256-2275
PBPI A2
AGGTGC TACA/hATTGAGAGG
2685-2666
PB P2 X I
CCAGGTTCCACTATGAAAGTG
1003-1023
PBP2X2
CATCCGTCAAACCGAAACGG
1294-I 275
PBP281
CAATCTAGAGTCTGCTATGGA
1636-I 656
PBP2B2
GGTCAATTCCTGTCGCAGTA
1712-I 693
430
PBP 2X (pbpBx) 292
PBP 2B (pbp2b) 77
Based on data in reference 8 (lytA), 3 (pbpla), 9 (pbp2x), and 10 (pbp2b).
191
J Infect Chemother 1997;3:190 197
vol); and the stacking gel modified to 4% (wt/vol) acrylamide, 0.063% N, N'-methylenebisacrylamide, 0.125 MTris-HC1 buffer (pH 6.8), 0.1% SDS, 0.1% T E M E D , and 0.1% APS (all reagents: Wako Pure Chemical, Tokyo, Japan). T h e voltages for SDS-PAGE were 60 V in the stacking gel and 90 V in the running gel.
Statistical analysis Statistical analysis was performed using Microsoft Excel Multivariate Analysis software (version 3.0). The partial regression coefficients, fixed values, and standard partial regression coefficients (Table 2) were obtained with multiple regression analysis.
RESULTS
Amplified DNA profile Figure 1 shows PCR-amplified DNA profiles obtained from 9 strains of S. pneumoniae.Three DNA bands, 430 bp in column A, and 292 bp and 77 bp in column B, correspond to the DNA fragments amplified with the primers that were designed from the sequences of the pbpla, pbp2x, and pbp2b genes of PSSP R6 strain. The 273 bp DNA fragment detected in column A correspond to the products of lytA gene for identification of S. pneumoniae species. When each DNA band ofpbpla, pbp2x, and pbp2b was detected, these PBP genes were regarded as having al-
Table 2. M u l t i p l e regression analysis between the presence of mutations of PBP genes and MICs of 9 fl-lactam antibiotics. Penicillin G
Ampicillin
Cefac]or Cefotaxime Cefpodoxime Cefdinir Cefditoren
Faropenem Imipenem
2.3135 1.0363 2.3359 -5.8835
2.1174 1.1839 2.3145 -5.2644
2.9277 1.1654 1.8175 -0.7587
1.4086 2.6813 0.5534 -5.9134
1.6985 2.9110 0.7959 -4,6045
2.0604 2.0003 1.0924 -3.8528
1.4863 2.7064 0.3658 -6.1469
1.6173 0.5339 2.2680 -6.5341
1.6480 0.5472 2.0861 -7.4817
0.4437 0.1841 0.4450
0.4144 0.2146 0.4500
0.5243 0.1933 0.3233
0.3348 0.5903 0.1307
0.3514 0.5578 0.1636
0.4283 0.3851 0.2255
0.3660 0.6174 0,0895
0.3784 0.1157 0.5270
0.3987 0.1226 0,5013
0.6066 0,4456 0.6231
0.5670 0.4889 0.6141
0.5907 0.3907 0.4253
0.3859 0.7573 0.1675
0.4163 0.7525 0.2166
0.4446 0.5744 0.2626
0.4520 0.8023 0.1279
0.4789 0.2537 0.6204
0.4950 0.2656 0.5977
0.8938 0.5767 0.8806
0.8831 0.5952 0.8741
0.8797 0.5751 0.8282
0.7234 0.8034 0.6423
0.7512 0.7922 0.6756
0.7959 0.6813 0.7301
0.7341 0.8287 0.6375
0.8625 0.5103 0.8818
0.8653 0.5165 0.8755
Coefficient of determination
0.8946
0.8870
0.8401
0.8003
0.8146
0.7679
0.8374
0.8501
0.8473
Multiple correlation co.*
0.8936
0.8859
0.8386
0.7984
0.8146
0.7657
0.8358
0.8486
0.8458
Partial regression coefficient
pbpla pbp2x pbp2b Fixed value Standard partial reg. coefficient
pbpla pbp2x pbp2b PartiaI correlation coefficient
pbp la pbp2x pbp2b Correlation coefficient
pbpla pbp2x pbp2b
* Multiple correlation coefficient adjusted for the degree of freedom.
Fig.1 Agarose gel electrophoresis of PCR-amplified D N A fragments of the lytA (273 bp) and pbpla (430-bp) genes in column A and pbp2x (292-bp) and pbp2b (77-bp) genes in column B in S. pneumoniae. These 3 PBP genes with the same sequences as penicillin-susceptible R6 strain 3,9,1~were amplified.
192
Fig.2 Fluorogram of the PBP of S. pneumoniae. M e m b r a n e fractions are solubilized after a 10 minute incubation with [3H]benzylpenicillin. MIC of benzylpenicillin and the results of PCR amplification (+: positive; - : negative) are listed b e l o w the illustration.
Rapid detection of Beta-lactam rosistance in S. pneumoniae
most the same sequences as those of the R6 strain of PSSE In contrast, when the D N A bands were not detected or different sizes were detected, their genes were regarded as possessing sequences different from those of the R6 strain.
Correctness of PCR detection of PBP genes To determine whether the presence or absence of mutations in the PBP genes had been correctly detected by PCR, the affinities of PBPs from the 25 strains selected at random were examined with [3H]benzylpeniciUin. The results for the 9 strains are shown in Fig. 2 together with the M I C of Pc and the results of P C R . W h e n the mutation(s) in the pbpla gene were detected, the rate of flow (Rf) value of PBP 1A from the mutationalpbpla was a little larger than that of the normal gene, and/or its affinity for [3H]benzylpenicitlin was decreased. PBP 2X from the mutatedpbp2x had a larger Rfvalue than did PBP 2A in this conditions of SDS-PAGE. When the mutation(s) was present in the pbp2b gene, PBP 2B was scarcely detected or its affinity for [3H]benzylpenicillin seemed to be reduced. From these findings, we concluded that the P C R using these sets of primers correctly reflects the changes of affinity in each PBE
Penicillin susceptibility and PBP genes Figure 3 shows the distribution of the M I C of Pc and the P C R results of pbpla, pbp2x, and pbp2b genes by P C R for the 310 strains of S. pneumoniae.
O f the 166 PSSP strains for which the M I C were _<0.063 pg/mL, 83 (50.0%) possessed normal genes of pbpla, pbp2x, and pbp2b. A m o n g the remaining 83 strains, 72 (43.4%) had a mutation in thepbp2x gene, and 1 (0.6%) in thepbp2b gene, 8 (4.8%) in thepbpla andpbp2x genes, and 2 (1.2%) in the pbp2x and pbp2b genes. O n the contrary, 104 (95.4%) o f the 109 P R S P strains for which the M I C was >_0.5 flg/mL had mutations in the three genes, pbpla, pbp2x, and pbp2b. The remaining 5 (4.6%) strains had a mutation in the pbpla and pbp2x gene. T h e 35 PISP strains for which the M I C ranged from 0.125 to 0.25 p g / m L possessed mutations in the pbpla, pbp2x, and pbp2b genes with various combinations, such as pbpla and pbp2b (11 strains), pbpla and pbp2x (4 strains), pbp2x and pbp2b (12 strains), etc.
Relationship between PBP ggnes and fl-lactam susceptibility Figure 4 shows the relationship between fl-lactam susceptibility and the pbpla, pbp2x, and pbp2b genes for the 310 pneumococcal strains. Two types were observed in these 8 fl-lactams. One is the Pc type by which the strains with pbp2x mutation were not separated from susceptible strains; ampicillin, c e f a c l o r , f a r o p e n e m , and imipenem belonged to this Pc type. A n o t h e r is the cefotaxime type by which the strains with pbp2x mutation were separated from susceptible strains with the
Fig.3 Correlation between MIC of benzylpenicillin and PBP gene mutations for a total 310 isolates of S. pneumoniae.
193
J Infect Chemother 1997;3:190- 797
Fig.4 Correlation between MIC of 8 ~-Iactam antibiotics and PBP gene mutations of 310 isolates of S. pneumoniae.
194
Rapid detection of Beta-lactam resistance in S. pneumoniae
M I C shifted to resistant side with 4 to 8 times higher values t h a n the f o r m e r ; cefdinir, c e f p o d o x i m e , and cefditoren belonged to this cefotaxime type.
Influence of PBP gene mutation on M I C level of fllactams Multiple regression analysis was applied to determine whether the presence of mutation of the 3 PBP genes influenced the M I C level of fl-lactams.Values of the M I C of fl-lactams for 310 p n e u m o c o c c a l strains changed to logarithm based 2 and the presence or absence of m u -
tation in pbpla, pbp2x, and pbp2b were used as a criterion variable and explanatory variable, respectively. Table 2 shows the partial regression coefficients, fixed values, and standard partial regression coefficients for the pbpla, pbp2x, and pbp2b obtained by the analysis. It also shows the coefficients of determination and multiple correlation coefficients adjusted for the degrees of freedom that indicate the accuracy of this analysis. Standard partial regression coefficients that indicate the weight of the explanatory variable showed high values forpbpla and pbp2b in Pc and ampicillin, for pbp2b in faropenem
Table 3. Estimated MIC value of ~ l a c t a m antibiotics calculated from multiple regression formula. Pattern of PBP gene mutation
Estimated MIC value (,u.g/mL) Cefaclor Cefotaxime Cefpodoxime Cefdinir C~fditoren
Penicillin G
Ampicillin
Normal
0.0169
0.0260
0.5910
0.0166
0.0411
0.0692
0.0141
Faropenem Imipenem 0.0141
0.0056
pbpla pbp2x pbp2b
0.0842 0.0347 0.0855
0.1129 0.0591 0.1294
4.4971 1.3257 2.0832
0.0440 0.1064 0.0243
0.1334 0.3062 0.0714
0.2887 0.2774 0.1476
0.0395 0.0921 0.0182
0.0433 0.0204 0.0680
0.0175 0.0082 0.0238
pbpla + pbp2x pbpla +pbp2b pbp2x + pbp2b
0.1727 0.4251 0.1754
0.2565 0.5616 0.2940
10.0867 5.8510 4.6725
0.2815 0.0646 0.1562
1.0035 0.2316 0.5368
1.1572 0.6156 0.5916
0.2581 0.0509 0.1187
0.0627 0.2085 0.0984
0.0256 0.0744 0.0347
pbpla+pbp2x+pbp2b
0.8719
1.2757
35.5530
0.4146
1.7422
2.4674
0.3325
0.3019
0.1088
Fig. 5 Predicted MIC values of fl-lactams calculated by inserting the variables of the ~ l a c t a m s and the PBP gene mutation into the multiple regression formula.
195
J Infect Chemother 1997;3:190-197
and imipenem, for pbp2x in cefotaxime, cefpodoxime, and cefditoren, and for pbpla in cefaclor and cefdinir. T h e value of multiple correlation coefficient adjusted for degrees of freedom ranged from 0.7657 for cefdinir to 0.8936 for Pc. Based on multiple regression analysis, the formula " M I C (log2) -- partial regression coefficient ofpbpla x mutation of pbpla + partial regression coefficient of pbp2x x mutation of pbp2x + partial regression coefficient ofpbp2b x mutation ofpbp2b - fixed value" holds good. Predicted M I C of/%lactams were calculated easily by inserting appropriate variables in the formula. Table 3 shows the M I C values estimated for 9 /3lactams that were calculated by entering the presence of mutations inpbpla, pbp2x, andpbp2b genes into the formula. These values are expected to be fit to the actual M I C value with a high probability when PBP gene mutation is demonstrated definitely by PCR. Figure 5 shows the predicted M I C of Pc, cefotaxime, cefaclor, and faropenem in cases of PBP gene mutation being observed in combinations. Nine antibiotics were classified into these 4 types. That is, the M I C of the Pc type including ampicillin were affected more strongly by the mutation ofpbpla and pbp2b than that ofpbp2x. In contrast, the M I C of the cefotaxime type including cefpodoxime and cefditoren were apparently more strongly affected by the mutation ofpbp2x than others. It was r e m a r k e d that the M I C o f f a r o p e n e m and imipenem were distinctly affected by pbp2b mutation, and those of cefaclor and cefdinir by pbpla mutation.
DISCUSSION T h e resistance ofS. pneumoniae to Pc and cefotaxime has been shown to be associated with mosaic m u t a tions in the pbpla, 11-13 pbp2b, 12-14 a n d pbp2x 7,15 genes. O f these genes, the m u t a t i o n s in the pbp2b mediate a low-level resistance to penicillin and pbp2x cefotaxime respectively. 6,16 A d d i n g pbpla m u t a t i o n s seems to cause a high resistance. 15 T h e s e findings suggest that rapid identification of P S S P and P I S P / P R S P m a y be possible by the detection of the m u t a tions in P B P genes. Pursuing this possibility, we investigated the relationship between penicillin susceptibility and the pbp2b gene. We found that 89.8% of the isolates for which the M I C was _> 1.0 p g / m L had a class B mutation of the pbp2b gene.1 F r o m these findings, it was suggested that the c o n c o m i t a n t detection of pbp2b and the lytA genes should allow us to identify simultaneously S. pneumoniae and PRSP. However, it was impossible to identify 10.2% of the isolates as PRSP, although the M I C of these isolates for penicillin was --_ 1.0 p g / m L and the affinity of PBP 2B was clearly decreased. Analysis of the sequences ofpbp2b genes in these strains showed that a 9-nucleotides repetition was inserted between the S W K and SSN of con-
196
served amino acid sequences in the transpeptidase region. 2 Many other types of variants were found. To raise the percentage of PRSP detection to compensate for this variety o f g e n e mutations, primers were designed to amplify the genes of susceptible strains but not of resistant strains. The clinical isolate was taken as PSSP when the D N A amplification occurred and as PISP or PRSP when it did not occur. The results clarified that mutations in the 3 PBP genes affected M I C . In addition, the multivariate analysis showed that the influence of PBP gene mutations on M I C differed according to the type of antibiotic. In other words, for penicillins, penem, and carbapenems, the pbpla and pbp2b gene mutations strongly affected M I C , while the pbp2x gene mutation had little influence. On the other hand, it was found that the pbp2x mutation h a d the strongest influence on the M I C o f m a n y cephems.The M I C of cefac~lor and cefdinir was strongly affected by the pbpla gene mutation. It is thus suggested that it is necessary to distinguish between penicillin resistance and cephem resistance for S. pneumoniae. This fact also means that the use of either a disc containing 1/lg ofoxacillin or a disc containing 2 units of Pc alone cannot distinguish correctly between the P C and cephem resistances. M I C for ]3-1actams can be predicted with a high probability by determining whether or not the pbpl a, pbp2x, and pbp2b gene mutations exist and. entering the results as explanatory variables in the multiple regression formula obtained by multivariate analysis. In clinical isolates assumed to be S. pneumoniae, examination of the 3 PBP genes together with lytA gene gives predicted values for susceptibility within 2 hours and is considered to be remarkably useful for the treatment of infectious diseases caused by & pneumoniae.
REFERENCES 1 Ubukata K, Asahi Y, Yamane A, Konno M. Combinational detection of autolysin and penicillin-binding protein 2B genes of Streptococcuspneumoniae by PCR. J Clin Microbiol 1996;34:592-596. 2 Yamane A, Nakano H, Asahi Y, Ubukata K, Konno M. Directly repeated insertion of 9-nucleotide sequence detected in penicillin-binding protein 2B gene of penicillin-resistant Streptococcuspneumoniae. Antimicrob Agents Chemother 1996;40:1257 1259. 3 Martin C, Sibold C, Hakenbeck R. Relatedness of penicillin-binding protein 1a genes from different clones of penicillin-resistant Streptococcus pneumoniae isolated in South Africa and Spain. EMBO J 1992; 11:3831-3836. 4 Dowson CG, Hutchison A, Brannigan JA, George RC, Hansman D, Lifiares J, et al. Horizontal transfer of penicillin-binding protein genes in penicillin-resistant clinical isolates of Streptococcus pneumoniae. Proc Natl Acad Sci USA 1989;86:8842-8846. 5 Dowson CG, Hutchison A, Spratt BG. Extensive re-modelling of the transpeptidase domain of penicillin-binding protein 2B of a penicillin-resistant South African isolate of Streptococcuspneumoniae. Mol Microbiol 1989;3:95-102.
Rapid detection of Beta-factam resistance in S. pneumoniae
6 Smith AM, Klugman KP. Alterations in penicillin-binding protein 2B from penicillin-resistant wild-type strains of Streptococcuspneumoniae. Antimicrob Agents Chemother 1995;39:859-867. 7 L a i b l e G , S p r a t t B G , H a k e n b e c k R. I n t e r s p e c i e s recombinational events during the evolution of altered PBP 2x genes in penicillin-resistant clinical isolates of Streptococcuspneumoniae. Mol Microbiol 1991 ;5:1993 2002. 8 Garcia P, Garcia JL, Garcia E, Lopez R. Nucleotide sequence and expression of the pneumococcal autolysin gene f r o m its own p r o m o t e r in Escherichia coli. G e n e 1986;43:265-272. 9 Laible G, Hakenbeck R, Sicard MA, Joris B, Ghuysen JM. Nucleotide sequences of the pbpX genes encoding the p e n i c i l l i n - b i n d i n g p r o t e i n s 2X from Streptococcus pneumoniae R6 and a cefotaxime-resistant mutant, C506. Mol Microbiol 1989;3:1337-1348. 10 Dowson CG, Hutchison A, Spratt BG. Nucleotide sequence of the penicillin-binding protein 2B gene of Streptococcus pneumoniae strain R6. N u c l e i c Acids Res 1989;17:7518. 11 Hakenbeck R, Ellerbrok H, Briese T, Handwerger S, Tomasz A. Penicillin-binding proteins of penicillin-sus-
12
13 14
15
16
ceptible and -resistant pneumococci: immunological relatedness of altered proteins and changes in peptides carrying the fl-lactam binding site. Antimicrob Agents Chemother 1986;30:553-558. Jabes D, Nachman S, Tomasz A. Penicillin-binding protein families: evidence for the clonal nature of penicillin resistance in clinical isolates ofpneumococci. J Infect Dis 1989;159:16-25. Markiewicz Z, Tomasz A. Variation in penicillin-binding protein patterns of penicillin-resistant clinical isolates of pneumococci. J Clin Microbiol 1989;27:405-410. Hakenbeck R, Tarpay M, Tomasz A. Multiple changes of penicillin-binding proteins in penicillin-resistant clinical isolates of Streptococcus pneumoniae. Antimicrob Agents Chemother 1980;17:364-371. Mufioz R, Dowson CG, Daniels M, CoffeyTJ, Martin C, Hakenbeck R, et al. Genetics of resistance to third-generation cephalosporins in clinical isolates of Streptococcus pneumoniae. Mol Microbiol 1992;6:2461-2465. GrebeT, Hakenbeck R. Penicillin-binding proteins 2b and 2x of Streptococcus pneuma~ae are primary resistance determinants for different classes of fl-lactam antibiotics. Antimicrob Agents Chemother 1996;40:829-834.
197