Detection of Macrolide Resistance in Mycoplasma pneumoniae by Real-Time PCR and High-Resolution Melt Analysis

Abstract

Mycoplasma pneumoniae is a significant cause of community-acquired pneumonia, which is often empirically treated with macrolides or azalides such as erythromycin or azithromycin. Recent studies have discovered the existence of macrolide-resistant strains within the population that have been mapped to mutations within the domain V region of the 23S rRNA gene. Currently, identification of these resistant strains relies on time-consuming and labor-intensive procedures such as restriction fragment length polymorphism, MIC studies, and sequence analysis. The current study reports two distinct real-time PCR assays that can detect the A2063G or A2064G base mutation (A2058G or A2059G by Escherichia coli numbering) conferring macrolide resistance. By subjecting the amplicon of the targeted domain V region of the 23S rRNA gene to a high-resolution melt curve analysis, macrolide-resistant strains can quickly be separated from susceptible strains. Utilizing this method, we screened 100 clinical isolates and found 5 strains to possess mutations conferring resistance. These findings were concordant with both sequencing and MIC data. This procedure was also used successfully to identify both susceptible and resistant genotypes in 23 patient specimens. These patient specimens tested positive for the presence of M. pneumoniae by a separate real-time PCR assay, although the bacteria could not be isolated by culture. This is the first report of a real-time PCR assay capable of detecting the dominant mutations that confer macrolide resistance on M. pneumoniae, and these assays may have utility in detecting resistant strains of other infectious agents. These assays may also allow for clinicians to select appropriate treatment options more rapidly and may provide a convenient method to conduct surveillance for genetic mutations conferring antibiotic resistance.

Mycoplasma pneumoniae is a human respiratory pathogen that is responsible for approximately 20% of all community-acquired pneumonia (14, 36, 39). Transmission among household contacts typically precedes the recognition of a community-wide outbreak, and the majority of exposed individuals may be asymptomatic or have mild illness and not seek care (2, 8). Nonetheless, severe disease involving multiple organs can occur and can lead to encephalitis and/or death (1, 5, 6, 32, 38). Recent outbreaks of M. pneumoniae infections have established that this agent is capable of causing a significant amount of morbidity and occasional mortality in the community (1, 18, 38). Previous studies have demonstrated that in outbreaks associated with closed settings, such as schools and hospitals, implementing appropriate antimicrobial prophylaxis treatment in a timely manner may shorten the course of the outbreak (12, 13). Because M. pneumoniae lacks a cell wall, infection is treated primarily with macrolide or azalide antibiotics such as erythromycin or azithromycin, although doxycycline and levofloxacin have also been used (4, 12, 13, 25). However, resistance to the macrolides and related azalides has been observed in M. pneumoniae and appears to have been increasing in recent years (17, 23, 24). A recent study found that nearly 20% of isolates from Japan between 2000 and 2003 were resistant to macrolide antibiotics (17). These data raise concerns that the current standard treatment for M. pneumoniae may be ineffective in certain instances. To our knowledge, there are no comprehensive studies documenting the incidence of macrolide-resistant M. pneumoniae in the United States. Such data could improve treatment recommendations, particularly during community-wide outbreaks and for individual cases requiring antibiotic intervention.

A few distinct single-base mutations in the peptidyltransferase (domain V) region of the 23S rRNA gene have previously been shown by MIC studies to confer high macrolide resistance (16, 17, 24). Current molecular detection methods for identifying these mutations rely on sequencing or restriction fragment length polymorphism analysis of domain V (17). Although both procedures are reliable, they are time-consuming and labor-intensive. A rapid and reliable assay to quickly determine macrolide resistance in M. pneumoniae would provide advantages over existing detection procedures. The current study reports two distinct real-time PCR assays followed by high-resolution melt (HRM) curve analysis for rapid detection of the presence of these mutations. LUX (light upon extension) and Sybr GreenER real-time PCR chemistries were used to screen 100 clinical isolates acquired over the last 18 years in order to examine the prevalence of macrolide resistance in M. pneumoniae. Lastly, we report that the LUX assay was capable of detecting these mutations in patient specimens for which an isolate was not recovered.

MATERIALS AND METHODS

Mycoplasma pneumoniae isolates.

Reference strains M129 (ATCC 29342D) and FH (ATCC 15531) were tested along with 100 clinical isolates submitted to or isolated within the Respiratory Disease Branch of the CDC for testing between 1991 and 2008. All isolates were grown in SP4 medium as previously described (33, 34). DNA was isolated from cultures and positive nasopharyngeal/oropharyngeal specimens associated with recent outbreaks by using the QIAamp DNA minikit (catalog no. 51304; Qiagen) according to the manufacturer's instructions.

Patient specimens.

The 30 oropharyngeal/nasopharyngeal swabs examined in this study were submitted to the CDC for M. pneumoniae testing. Confirmation was determined by serological and real-time PCR assays. Serological tests were conducted using the Immunocard Mycoplasma IgM kit (catalog no. 709030; Meridian Bioscience) and the Mycoplasma pneumoniae IgG/IgM antibody test (catalog no. R24229; Remel). Real-time PCR testing was performed with M. pneumoniae-specific markers as previously described (41).

Real-time PCR.

Primers for the LUX and Sybr GreenER real-time PCR assays were designed to amplify the domain V region of the 23S rRNA gene of M. pneumoniae (Table 1). Sybr assay primers were manually designed to selectively amplify a 112-bp amplicon spanning the region conferring resistance. D-LUX design software (Invitrogen) was used to generate a 6-carboxyfluorescein-labeled forward primer. The reverse primer was manually designed and was used along with the labeled forward primer to amplify a 158-bp amplicon of the target region.

TABLE 1.

Primers used for real-time PCR amplification and sequencing^a

Procedure	Name	Primer sequence (5′→3′)	Amplicon size (bp)
Sequencing	Forward_1937	AACTATAACGGTCCTAAGGTAGCG	217
	Reverse	GCTCCTACCTATTCTCTACATGAT
LUX amplification	Forward labeledb	gacagtcTGGTGTAACCATCTCTTGACTG“t”C	158
	Reverse	GCTCCTACCTATTCTCTACATGAT
Sybr GreenER amplification	Forward_2020	TCCAGGTACGGGTGAAGACA	112
	Reverse	GCTCCTACCTATTCTCTACATGAT

^aThe reverse primer is the same for all three PCR assays. All three sets of PCR primers amplify the domain V region of the 23S rRNA gene.

^b6-Carboxyfluorescein is attached to the “t” on the 3′ end of the forward primer. The lowercased bases on the 5′ end are added by the software to create a quenching hairpin formation when the primer is in the unbound state.

The Sybr reaction mixture was prepared using the Universal Sybr GreenER qPCR kit (catalog no. 11762-100; Invitrogen), containing the following components per reaction: 12.5 μl of 2× master mix, 100 nM (final concentration) forward and reverse primers, 5 ng of the template, and nuclease-free water (catalog no. P1193; Promega) to a total reaction volume of 25 μl. Amplification was performed on the Rotor-Gene 6000 system (catalog no. 65H0; Corbett) under the following conditions: 1 cycle of 50°C for 2 min and 1 cycle of 95°C for 10 min, followed by 45 cycles of 95°C for 15 s and 60°C for 60 s, with data acquired on the 60°C step in the green channel. Following amplification, HRM was performed between 74°C and 82°C at a rate of 0.03°C per step.

The LUX PCR mixture was prepared using the Platinum Quantitative PCR SuperMix-UDG kit (catalog no. 11730-025; Invitrogen), containing the following components per reaction: 12.5 μl of 2× master mix, 125 nM (final concentration) forward and reverse primers, 0.25 μl of Platinum Taq polymerase (5 U per μl), 5 ng of the template, and nuclease-free water (catalog no. P1193; Promega) to a final volume of 25 μl. Real-time PCR amplification was performed using the Corbett Rotor-Gene 6000 system under the following cycling conditions: 1 cycle of 95°C for 2 min, followed by 45 cycles of 95°C for 15 s, 60°C for 30 s, and 72°C for 30 s, with data acquired on the 72°C step in the green channel. Following amplification, HRM was performed between 79°C and 86°C at a rate of 0.02°C per step. All specimens were tested in triplicate.

HRM data analysis.

The HRM curves are derived by first selecting two normalization regions, one occurring prior to the melting of the double-stranded product and one following complete separation of the two strands. Each region is generated by default by the software but may be manipulated manually to achieve optimum results. These regions function to normalize the fluorescence of the melt curves from the raw channels by averaging all starting and ending fluorescence values such that the end point value of each sample is identical to the average. This allows for the melting curve profile of each isolate to be analyzed relative to those of the others. For the Sybr GreenER assay, region one was between 76.5°C and 77°C, while region two was between 80.5°C and 81°C. The LUX assay had normalizing region one between 81°C and 81.5°C and region two between 83.5°C and 84°C.

The “difference graph” was generated by selecting the melting profile of the M129 reference strain as the susceptible genotype. All susceptible genotypes were then normalized to zero, and any deviations (i.e., resistant isolates) from this reference genotype were highlighted in the difference graph as a positive curve.

Sequencing.

Domain V of the 23S rRNA gene (accession no. X68422) was amplified using primers 5′-AACTATAACGGTCCTAAGGTAGCG-3′ and 5′-GCTCCTACCTATTCTCTACATGAT-3′. The PCR mixture was prepared using the Platinum Quantitative PCR SuperMix-UDG kit, containing the following components per reaction: 12.5 μl of 2× mix, 100 nM (final concentration) each primer, 0.25 μl of Platinum Taq polymerase (5 U per μl), 5 ng of the template, and nuclease-free water (catalog no. P1193; Promega) to a final volume of 25 μl. Amplification was performed using the DNA Engine Dyad Peltier thermocycler (catalog no. PTC-0220G; Bio-Rad) under the following conditions: 1 cycle of 95°C for 2 min, followed by 45 cycles of 95°C for 15 s, 60°C for 30 s, and 72°C for 60 s. The 200-bp PCR product was purified using the Geneclean Turbo kit (catalog no. 1102-600; qBiogene) according to the manufacturer's instructions. The sequencing reaction was performed on each strand of the amplicon with the BigDye Terminator cycle sequencing kit (version 3.1; catalog no. 4337454; Applied Biosystems Inc.) by using the thermocycler described above under the following conditions: 1 cycle of 96°C for 1 min, followed by 35 cycles of 96°C for 10 s, 55°C for 10 s, and 60°C for 3 min. The product was purified using CentriSep 8 spin columns (catalog no. CS-912; Princeton Separations) according to the manufacturer's instructions, followed by sequencing on an ABI 3130XL instrument (Applied Biosystems Inc.) using standard conditions for a 50-cm capillary. Consensus sequences were obtained using DNAStar Lasergene SeqMan Pro software and were aligned with ClustalW software.

MIC determination.

MICs of erythromycin (catalog no. E5389-5G; Sigma), tetracycline (catalog no. T7660-5G; Sigma), and minocycline (catalog no. M9511-250MG; Sigma) were determined for each resistant genotype and two susceptible isolates, along with the M129 and FH reference strains, by an agar dilution method as previously described (37). SP4 agar plates were prepared using twofold dilutions of each of these antibiotics, from a starting concentration of 256 μg/ml down to 0.008 μg/ml, as well as with no antibiotic. All isolates were tested at two dilutions (10⁴ and 10⁵ CFU/ml), and the plates were incubated at 37°C for a total of 21 days. The MIC end point was determined as the lowest concentration of the antimicrobial agent at which the growth of the organism was inhibited while the antimicrobial-free control plate demonstrated growth of the organism.

RESULTS

Real-time PCR and HRM analysis.

Both the LUX and Sybr GreenER real-time PCR assays were used to screen M. pneumoniae reference strains M129 and FH, along with 100 clinical isolates acquired by the Respiratory Disease Branch of the CDC from 1991 to 2008. After each real-time PCR assay and HRM analysis were performed on the 100 clinical isolates, 5 were found to possess single-base mutations that have previously been shown to confer macrolide resistance. Table 2 lists these five isolates along with their respective origins and years of isolation.

TABLE 2.

Macrolide-resistant clinical isolates and MIC data

Isolate	Yr of isolation	Mutationa	Location	MIC (μg/ml) of the following antibiotic:
				Erythromycin	Tetracycline	Minocycline
M129	N/A	N/A	ATCC	0.063	0.250	0.250
FH	N/A	N/A	ATCC	0.063	0.250	0.250
6	2007	A2064G	Rhode Island	128	0.250	0.125
11	2007	A2064G	Rhode Island	128	0.250	0.125
2p	2007	A2064G	Rhode Island	128	0.250	0.250
18	2007	N/A	Rhode Island	0.063	0.125	0.125
19	2007	N/A	Rhode Island	0.063	0.250	0.125
685	Unknown	A2063G	Denmark	128	0.250	0.125
1006	1998	A2063G	New York	128	0.125	0.125

^aPositions are based on M. pneumoniae numbering.

The results of the HRM analysis of the Sybr GreenER assay are shown in Fig. 1. The Sybr GreenER assay exhibits a melting pattern that begins its separation at approximately 77°C and ends at approximately 80.5°C (Fig. 1A). The reference strains and susceptible isolates show a lower-temperature melting profile, resulting in the separation of this group from the resistant isolates. Resistant isolates maintain a higher level of fluorescence for a longer time than susceptible isolates, resulting in a melting profile that is shifted to the right. Figure 1B displays the same data using a “difference graph,” which shows a clear separation between the resistant and susceptible isolates. This is accomplished by defining a reference strain as a susceptible genotype. The fluorescence level of the susceptible isolates is normalized to zero, and any deviations from this standard are recorded in the difference graph.

FIG. 1.

Sybr GreenER assay isolate data. (A) High-resolution melting profile in the normalized graph mode. The florescence level is normalized to 100 for all isolates so that the slope and shifts of each isolate can be easily compared. (B) Genotype difference graph. The susceptible isolates have been normalized such that any deviations from their patterns can be observed.

The results of the HRM analysis of the LUX assay are shown in Fig. 2. The LUX assay exhibits a melting pattern that begins its separation at 81.5°C and ends at approximately 83.5°C (Fig. 2A). The susceptible isolates exhibit an earlier decrease in fluorescence level than the resistant isolates. This results in a melting profile that shows the resistant isolates being shifted to the right on the graph compared to the susceptible isolates. Similarly, the difference graph shows a clear distinction between the resistant and susceptible isolates (Fig. 2B).

FIG. 2.

LUX assay isolate and patient specimen data. (A) High-resolution melting profile in the normalized graph mode; (B) genotype difference graph.

For each chemistry tested, the macrolide-resistant isolates exhibited a highly reproducible (all specimens tested in triplicate) melting profile distinct from that of the macrolide-susceptible isolates (including reference strains). The LUX assay demonstrates a slightly higher HRM temperature range than the Sybr GreenER assay; however, both assays demonstrate reliable and obvious separation of macrolide-resistant strains from macrolide-susceptible strains.

Patient specimen testing.

Patient specimens from 30 PCR- and serology-confirmed M. pneumoniae cases were tested using the LUX assay, and the results are shown in Table 3. Seven patient specimens failed to amplify, most likely due to limited sensitivity of the assay. Five patient specimens from 2006 and 2007 tested positive for resistance; three of them yielded isolates (isolates 6, 11, and 2p) that were confirmed to be resistant by sequence analysis and MIC studies. Susceptibility was seen in 18 specimens; 10 isolates were obtained and confirmed to be susceptible by sequence analysis (Fig. 3 and Table 3). For those specimens from which no isolate was recovered (n = 10), amplification and melting patterns consistent with the resistant or susceptible genotype shown in Fig. 2 were observed. The Sybr GreenER assay proved to be unreliable for testing patient specimens, as discussed below.

TABLE 3.

Patient specimen information^a

Patient specimen no.	Corresponding isolate	PCR predictionb		Sequencing resultc
		Specimen (LUX assay)	Isolate (LUX and SYBR assays)
1	6	R	R	A2064G mutation
2	11	R	R	A2064G mutation
3	2P	R	R	A2064G mutation
4	3	S	S	Wild type
5	10	S	S	Wild type
6	13	S	S	Wild type
7	15	S	S	Wild type
8	17	S	S	Wild type
9	18	S	S	Wild type
10	19	S	S	Wild type
11	20	S	S	Wild type
12	3064	S	S	Wild type
13	3078	S	S	Wild type
14	None	R	N/A	N/A
15	None	R	N/A	N/A
16	None	S	N/A	N/A
17	None	S	N/A	N/A
18	None	S	N/A	N/A
19	None	S	N/A	N/A
20	None	S	N/A	N/A
21	None	S	N/A	N/A
22	None	S	N/A	N/A
23	None	S	N/A	N/A
24	None	No amplification	N/A	N/A
25	None	No amplification	N/A	N/A
26	None	No amplification	N/A	N/A
27	None	No amplification	N/A	N/A
28	None	No amplification	N/A	N/A
29	None	No amplification	N/A	N/A
30	None	No amplification	N/A	N/A

^aAll 30 patient specimens were confirmed to be positive for M. pneumoniae by using separate real-time PCR and serological assays. Isolates were obtained from 13 of 30 specimens. Specimens from which no isolate was obtained had insufficient volume for sequencing.

^bR, resistant; S, susceptible; N/A, not applicable.

^cMutation positions are based on M. pneumoniae numbering.

FIG. 3.

Sequencing and alignment data for a 46-bp region of the 200-bp amplicon of the domain V region of the 23S rRNA gene. Base pairs 2035 through 2080 are shown according to M. pneumoniae numbering. Base positions 2063 and 2064 are underlined. A dash indicates that the base is identical to that in the two reference strains and the GenBank entry (accession no. X68422).

Sequencing analysis.

The domain V region of the 23S rRNA gene was sequenced for reference strains M129 and FH along with 29 susceptible and 5 resistant isolates. Of the five PCR-predicted resistant isolates, three contained the A2064G transition and two contained the A2063G transition (Table 2). The sequencing and alignment data for the region spanning the mutations are shown in Fig. 3; they demonstrated 100% correlation with the results predicted by the real-time PCR assays.

Antimicrobial susceptibility.

The five isolates that possessed the A2063G or A2064G mutation demonstrated a 2,048-fold higher erythromycin MIC than the reference strains and the isolates not containing the mutation (Table 2). All isolates and the reference strains had low MICs for both tetracycline and minocycline, as predicted. The positive-control plates exhibited the expected titers for the reference strains and the isolates (data not shown).

DISCUSSION

Community-acquired pneumonia caused by M. pneumoniae is commonly treated with macrolides or azalides such as erythromycin, clarithromycin, or azithromycin. Alternatives such as tetracyclines and fluoroquinolones may be used but are not recommended for children due to concern about possible adverse effects. Recent reports of macrolide resistance have prompted greater scrutiny of treatment options and epidemiological monitoring of M. pneumoniae infections in different parts of the world (17, 19, 23, 24). Indeed, recent studies in Japan reported that no resistant M. pneumoniae isolates were detected between 1985 and 1999. However, more than 17% of the isolates tested from 2000 to 2006 were resistant to a variety of macrolides, including erythromycin, azithromycin, clarithromycin, and josamycin (17, 23). While clinical isolates from France, Finland, and the United States have also been reported to be resistant to macrolides on occasion, to date few data exist describing the prevalence of macrolide resistance in M. pneumoniae in the United States (4, 24). Although it is well established that point mutations of A2063G or A2064G within domain V of the 23S rRNA gene account for macrolide resistance, no rapid diagnostic assays exist for detecting this change.

We developed two distinct real-time PCR assays that can readily detect the mutations responsible for macrolide resistance in M. pneumoniae. By performing HRM analysis on isolate amplicons, we were able to rapidly distinguish between the susceptible isolates and the isolates that were resistant due to the A-to-G mutation. This transition causes a distinct difference in the thermodynamic properties of the amplicon, leading to the real-time detection of the resistant genotype. The amplicon containing the G base causes the product to melt at a slightly higher temperature, producing the noticeable shift in the melting temperature displayed in Fig. 1 and 2. This was consistent in both the LUX and Sybr GreenER assays, and the position of the mutation (A2063G or A2064G), had no impact on the results. The ability to detect these single-base mutations is due to the extraordinary discriminatory power of the advanced HRM technology in conjunction with superior detection chemistries. HRM is achieved by measuring the fluorescence of an amplicon as the temperature is slowly increased, providing greater sensitivity than traditional melt curves because of the enhanced thermal resolution and the ability to acquire as many as 1,000 data points for each degree Celsius change (40). This increased-resolution methodology has also been used to detect single-nucleotide polymorphisms in human genes (9, 11, 15, 40, 42).

Intercalating dyes such as Sybr green (and derivatives) are often used with HRM analysis; likewise, labeled primers have also been used to detect single-nucleotide polymorphisms (9). LUX chemistry also affords the ability to perform melt curve analysis but has yet to be tested on an HRM platform. LUX chemistry utilizes a fluorescent molecule attached to the primer on the 3′ end, and quenching is accomplished by the presence of 6 to 7 bases attached to the 5′ end. These bases are complementary to the 3′ end, creating a hairpin structure when the primer is in an unbound state, thus eliminating the need for a quencher molecule (3). During amplification, the florescent molecule is incorporated into the amplicon, thus allowing for HRM analysis following amplification (3). Our results shown in Fig. 2 indicate that LUX chemistry can be used reliably in HRM analysis to detect macrolide resistance in M. pneumoniae, although the separation observed is not as large as that seen in the Sybr GreenER assay (Fig. 1). This may be caused by a few factors. First, it has been observed that shorter amplicons are more sensitive for detection of minor changes and offer higher resolution (9, 15). The amplicon generated during LUX amplification is 158 bp, while the Sybr GreenER amplicon is 112 bp. This difference of 46 bp could certainly account for the difference in separation between the two assays. A second factor is the intercalating nature of Sybr green molecules, which are present throughout the amplicon, leading to a greater discrepancy in fluorescence between the resistant and sensitive genotypes. Nonetheless, LUX chemistry provides a specific, convenient, and economical means for rapid detection of a single-base mutation in this region.

Examination of our isolates collected in the United States over the past 18 years showed no macrolide resistance prior to 1998 (n = 53). One isolate submitted from Denmark in 1996 (isolation date unknown) did have macrolide resistance. Interestingly, in 1998 an isolate, designated 1005, was obtained from a patient during an outbreak in New York. This isolate demonstrated a susceptible genotype (Fig. 3). Three weeks later, another isolate (designated 1006) was obtained from the same patient, who continued to have symptoms. This second isolate was shown to possess macrolide resistance through acquisition of the A2063G mutation (Table 2). A comparison of the sequence data in Fig. 3 shows the resistance mutation (A2063G) for isolate 1006, while isolate 1005 displays the susceptible genotype of M. pneumoniae. The evolution of a strain from a susceptible to a resistant genotype within a patient receiving macrolide therapy has been previously documented (20, 21, 27, 28, 31). This phenomenon highlights the importance of having a rapid diagnostic assay for determining macrolide resistance.

A recent M. pneumoniae outbreak in the United States (2006 to 2007) yielded three isolates from separate patients that possessed the A2064G mutation. To our knowledge, this is the first and only instance of macrolide resistance caused by this specific mutation in isolates from the United States. The significance of this finding is unclear, but a prior MIC study showed that the A2064G mutation causes a higher and broader resistance to macrolides than the A2063G mutation, suggesting that strains possessing the A2064G mutation may be more difficult to treat than those with the A2063G mutation (17). While the total number of isolates recovered from this recent U.S. outbreak was minimal (n = 11), 27% were macrolide resistant, in comparison to 2 resistant isolates identified among the 89 isolates tested (2.2% resistance) prior to this outbreak. If substantiated by broader testing, this trend would be similar to what has recently been observed in Japan (17).

In the current study, the macrolide erythromycin was chosen for MIC experiments to demonstrate the correlation between resistant genotypes and the level of resistance. The MIC for each macrolide-resistant isolate was 128 μg/ml, while that for susceptible strains was 0.063 μg/ml, with all isolates remaining sensitive to tetracycline and minocycline as expected (Table 2). All tests (both the LUX and Sybr GreenER real-time PCR assays, amplicon sequencing, and MIC studies) demonstrated 100% correlation for all of the isolates tested. Furthermore, testing of patient specimens using the LUX assay also showed complete correlation with the corresponding isolates; however, the Sybr assay was unreliable for this purpose. This is likely due to the fact that Sybr green is incorporated into any double-stranded DNA present in the sample, and since patient clinical specimens undoubtedly contain large amounts of nonspecific DNA, HRM curves display variable melt profiles that are not reproducible. In contrast, LUX amplification produces fluorescence only when the fluorescent molecule is incorporated within the specific targeted amplicon and displays no fluorescence in the presence of nonspecific, contaminating DNA that may be contained in the sample. This allows for consistent HRM analysis of clinical specimens by use of LUX chemistry (Fig. 2).

In summary, this report describes two separate real-time PCR assays that are capable of determining macrolide resistance in M. pneumoniae from isolates, one of which, the LUX assay, also enables HRM analysis of patient specimens. Additionally, we report the existence of macrolide-resistant strains, containing both A2063G and A2064G genotypes, in the United States. These findings suggest that some patients with M. pneumoniae infection may require alternative therapy, especially when little or no improvement is observed after administration of macrolides or azalides. This study also underscores the need for surveillance of macrolide resistance in M. pneumoniae, particularly in community-based outbreaks. The current assays may also have utility for resistance testing of other clinically relevant agents, because the mutations described here confer high-level resistance to macrolide antibiotics on Helicobacter pylori, Mycobacterium spp., E. coli, Streptococcus pneumoniae, and other organisms (22, 29, 30, 35). Previous studies have also documented that bacterial strains that acquire macrolide resistance may simultaneously develop resistance to lincosamides, ketolides, and group B streptogramins (7, 10, 26). The real-time PCR/HRM assays described here may be adapted to detect these changes as well. The assays reported in the current study provide a vast improvement in the diagnostic ability to detect macrolide resistance in M. pneumoniae and document the presence of these strains within the United States. These results warrant further studies to characterize and determine the prevalence of macrolide-resistant M. pneumoniae strains within population groups.