Alterations of Penicillin-Binding Proteins 1A, 2X, and 2B in Streptococcus pneumoniae Isolates for Which Amoxicillin MICs Are Higher than Penicillin MICs

Abstract

Penicillin-binding proteins (PBPs) of 15 selected penicillin- and amoxicillin-resistant Streptococcus pneumoniae isolates (MICs of 2 to 8 and 8 to 16 μg/ml, respectively) were studied. In addition to typical changes in PBPs 1A and 2X, these strains had 10 unique changes in PBP 2B, including a ₆₁₈A-G substitution, which may be the key alteration associated with amoxicillin resistance.

The efficacy of β-lactam antibiotics has been compromised by rapid increases in the prevalence of penicillin-nonsusceptible Streptococcus pneumoniae strains (1). This is the result of altered penicillin-binding proteins (PBPs) with decreased antibiotic affinities (2-4, 12, 15, 17). Pneumococci have, however, generally remained susceptible to amoxicillin as a result of the favorable pharmacokinetic properties of this agent (6, 13, 14). A few reports have documented isolates that are resistant to both amoxicillin and penicillin, with, in some instances, the unusual feature of amoxicillin MICs (4 to 16 μg/ml) being higher than penicillin MICs (2 to 8 μg/ml) (9, 10, 15). The purpose of this study was to characterize the nucleotide sequences of regions encoding the penicillin-binding domains of pbp1a, pbp2b, and pbp2x from representative strains with this unusual pattern.

To accomplish this purpose, 15 such isolates of S. pneumoniae were selected from the 1998 to 2000 Alexander Network and Augmentin XR 2000 surveillance studies and the MICs for these isolates were confirmed by broth microdilution (Trek Diagnostic Systems, Inc., Cleveland, Ohio) (16). Ten other penicillin-resistant but amoxicillin-susceptible strains were also selected for comparative purposes.

Analysis of nucleotide and deduced amino acid sequences encoding penicillin-binding domains of PBPs was performed by PCR using a standard protocol (15), with the addition of primers to amplify an additional fragment of the pbp1a gene, 1AAMPF (5′ ACCTACTCACAACTGGGATGGATG) and 1AAMPR (5′ TGGTTGTGCTGGTTGAGGATTCTG), which correspond to positions 869 and 892 and are complementary to positions 2134 and 2157, respectively. The PCR products produced were a 1.2-kb fragment of the pbp1a gene, a 2-kb segment of the pbp2x gene, and a 1.4-kb segment of the pbp2b gene. PCR products were purified from primers and excess nucleotides with a QIAquick PCR purification kit (QIAGEN, Valencia, Calif.) and directly sequenced with a CEQ8000 genetic analysis system (Beckman Coulter, Fullerton, Calif.). DNA sequences were aligned with ClustalW (18) and compared to nucleotide and deduced amino acid sequences from penicillin-susceptible S. pneumoniae R6. Mutants with nucleotide changes in a particular gene were sequenced twice forward and twice in the reverse direction on PCR products amplified in independent reactions. Strains were also serotyped with pneumococcal antisera (Statens Serum Institute, Copenhagen, Denmark), and DNA fingerprinting was determined by pulsed-field gel electrophoresis (PFGE) after digestion of DNA with SmaI (8).

The surveillance studies from which the strains were selected included 16,141 isolates of S. pneumoniae from 29 countries. The results for penicillin/amoxicillin susceptibility were 64.9/93.6% susceptible, 13.9/3.2% intermediate, and 21.1/3.2% resistant. The amoxicillin MICs for a total of 483 isolates (3.0%) of 8 to 16 μg/ml that were 1 or more dilutions higher than penicillin MICs were predominantly from the United States (381 of 4,928 isolates; 7.7%), Spain (29 of 537 isolates; 5.4%), and France (19 of 821 isolates; 2.3%).

Penicillin and amoxicillin MICs for 23 of the 25 isolates were confirmed to be within 1 doubling dilution of initial MICs, while the amoxicillin MICs for two strains were 2 doubling dilutions lower. The origins, serotypes, and susceptibility patterns of the 15 penicillin- and amoxicillin-resistant isolates are shown in Table 1. PFGE profiles showed 14 unique patterns, with isolates 8 and 31 sharing a common profile.

TABLE 1.

Origins, serotypes, overall susceptibility patterns (initial MICs), and PFGE profiles of 15 S. pneumoniae isolates with 10 common substitutions in PBP 2B (grouped according to strains with similar MIC patterns)^a

Isolate no.	Origin	Yr iso- lated	Speci- men source	Patient age	Sero- type	MIC for isolate (μg/ml)
						PEN	AMX	AMC	CRO	AZM	CLR	CLI	CPR	DOX	ERY	CEC	CFM	LVX	SXT
1	New York, NY	2000	SP	37 yr	19F	4	16	16	1	>32	>32	>1	>16	8	>32	>16	8	1	8
9	Cleveland, Ohio	1999	SP	60 yr	19A	4	16	16	4	>32	>32	>1	>16	8	>32	>16	>16	1	8
10	Madrid, Spain	1999	EA	C	6B	4	16	16	1	>32	>32	>1	>16	4	>32	>16	8	0.5	4
15	Madison, Wis.	2000	EA	9 mo	14	8	16	8	4	>32	>32	>1	>16	4	>32	>16	>16	1	8
25	Betera, Spain	1999	EA	A	14	8	16	16	2	>32	>32	>1	16	8	>32	>16	16	0.5	4
29	Betera, Spain	1999	SP	C	6B	4	16	16	1	>32	>32	>1	>16	8	>32	>16	8	1	8
2	Mobile, Ala.	2000	SP	19 mo	14	4	8	16	2	8	4	0.06	>16	0.25	8	>16	8	1	8
16	Fargo, N.D.	2000	NP	18 mo	14	8	8	16	8	4	4	0.06	>16	0.12	4	>16	>16	0.5	8
17	Cleveland, Ohio	2000	BL	2 yr	14	8	16	8	4	8	4	0.12	>16	0.25	8	>16	>16	0.5	8
20	Cleveland, Ohio	2000	BL	70 yr	6B	4	16	16	1	8	8	0.12	>16	0.25	8	>16	8	0.5	8
8	Charlottosville, Va.	2000	SI	99 yr	6B	2	8	8	1	0.06	0.03	0.12	16	>8	0.06	>16	4	0.5	4
31	Madison, Wis.	2000	SP	19 yr	6B	2	8	8	2	0.06	0.03	0.06	>16	8	0.06	>16	16	1	4
22	Indianapolis, Ind.	2000	BA	31 yr	23F	4	8	8	2	2	2	0.06	>16	8	4	>16	8	1	4
14	Cleveland, Ohio	2000	SI	3 yr	23F	8	16	16	4	0.06	0.03	0.12	>16	0.25	0.06	>16	>16	0.5	4
23	Mexico City, Mexico	1999	SP	72 yr	19F	4	16	16	4	4	2	0.12	>16	0.25	4	>16	16	1	1

^aPFGE showed that all isolates had unique patterns; except isolates 8 and 31, which had similar PFGE patterns, PEN, penicillin; AMX, amoxicillin; AMC, amoxicillin/clavulanic acid; CRO, ceftriaxone; AZM, azithromycin; CLR, clarithromycin; CLI, clindamycin; CPR, cefprozil; DOX, doxycycline; ERY, erythromycin; CEC, cefeclor; CFM, cefuroxime; LVX, levofloxacin; SXT, trimethoprim/sulfamethoxazole; SP, sputum; EA, ear; NP, nasopharyngeal; BL, blood; SI, sinus; BA, bronchoalveolar lavage; C, child; A, adult.

PBP analyses showed that all but two of the 25 strains had T₃₇₁A or S substitutions in the STMK motif of PBP 1A. There were 14 strains with a T₃₇₁S substitution in PBP 1A, the penicillin MICs for 12 of which were lower than the amoxicillin MICs. Within the ₃₃₇STMK₃₄₀ motif of PBP 2X, there were six strains with an M₃₃₉F substitution, the penicillin MICs for all of which were lower than the amoxicillin MICs. Another substitution within the ₃₃₇STMK₃₄₀ motif of PBP 2X was T₃₃₈S or A, which occurred in all but one of the strains tested. All 25 strains had a T₄₄₅A substitution in the ₄₄₂SSNT₄₄₅ motif of PBP 2B (Table 2).

TABLE 2.

Amino acid alterations of the three conserved motifs of PBP 1A, 2X, and 2B in S. pneumoniae R6 and 25 study strains

Strain no.	Changes in amino acids of conserved motifs forming or surrounding active PBP binding site
	PBP 1A			PBP 2X			PBP 2B
	370-373a	428-430	557-559	337-340	394-397	546-549	385-388	442-445	614-616	No. of changes in 590-641 region of PBP 2B
R6	STMKb	SRN	KTG	STMK	HSSN	LKSG	SVVK	SSNT	KTG	0
11	SAMK	—	—	SAMK	—	VKSG	—	SSNA	—	0e
21	SAMK	—	—	SAMK	—	VKSG	—	SSNA	—	0e
30	SAMK	—	—	SAMK	—	—	—	SSNA	—	0e
26	SAMK	—	—	SAMK	—	VKSG	—	SSNA	—	1e
3	SAMK	—	—	SAMK	—	VKSG	—	SSNA	—	2e
24	SSMK	—	—	SAMK	—	VKSG	—	SSNA	—	3e
27	SAMK	—	—	SAMK	—	VKSG	—	SSNA	—	4e
32	SSMK	—	—	SAMK	—	—	—	SSNA	—	4e
33	SSMK	—	—	SAMK	—	—	—	SSNA	—	4e
6	—d	—	—	—	YSSN	—	—	SSNA	—	7e
1	SSMKc	—	—	SSMK	—	VKSG	—	SSNA	—	10f
2	SSMK	—	—	SAMK	—	VKSG	—	SSNA	—	10f
8	SAMK	—	—	SAMK	—	VKSG	—	SSNA	—	10f
9	SSMK	—	—	SSMK	—	VKSG	—	SSNA	—	10f
10	SSMK	—	—	SAMK	—	VKSG	—	SSNA	—	10f
14	SSMK	—	—	SAFK	—	VKSG	—	SSNA	—	10f
15	SSMK	—	—	SAFK	—	VKSG	—	SSNA	—	10f
16	SSMK	—	—	SAFK	—	VKSG	—	SSNA	—	10f
17	SSMK	—	—	SAFK	—	VKSG	—	SSNA	—	10f
20	—	—	—	SAMK	—	VKSG	—	SSNA	—	10f
22	SSMK	—	—	SAMK	—	VKSG	—	SSNA	—	10f
23	SSMK	—	—	SFFK	—	VKSG	—	SSNA	—	10f
25	SAMK	—	—	SAMK	—	VKSG	—	SSNA	—	10f
29	SSMK	—	—	SAMK	—	VKSG	—	SSNA	—	10f
31	SAMK	—	—	SAFK	—	VKSG	—	SSNA	—	10f

^aAmino acid position numbers of amino acids shown in column below.

^bFor S. pneumoniae R6, conserved amino acid motifs are shown in boldface and flanking amino acids (where applicable) are shown in regular typeface.

^cChanges in amino acid motifs of test strains from those of S. pneumoniae R6 are shown in boldface for study strains.

^d—, no change in amino acid motifs of test strains from those of S. pneumoniae R6.

^eChanges included one or more of the following substitutions: 608L→A, 624D→G, 627 Q→E and 629T→N.

^fChanges included 591A→S, 596G→P, 605N→D, 608L→T, 618A→G, 624D→G, 627 Q→E, 629T→N, 639S→T, and 640D→E in all of these strains.

Based on reports that the 590- to 641-amino-acid region of PBP 2B, which surrounds the ₆₁₄KTG₆₁₆ drug binding site motif, is associated with amoxicillin resistance (10, 12), all of our strains were also analyzed for alterations in this region. Strains were categorized based on the number of amino acid substitutions surrounding the ₆₁₄KTG₆₁₆ motif of PBP 2B. In addition to changes present in amoxicillin-susceptible, penicillin-resistant, strains, a pattern of 10 substitutions (A₅₉₁S, G₅₉₆P, N₆₀₅D, L₆₀₈T, A₆₁₈G, D₆₂₄G, Q₆₂₇E, T₆₂₉N, S₆₃₉T, and D₆₄₀E) was found to be unique to the 15 amoxicillin-resistant isolates, with 6 substitutions being specific for this group: G₅₉₆P, N₆₀₅D, L₆₀₈T, A₆₁₈G, S₆₃₉T, and D₆₄₀E. One of these unique alterations, A₆₁₈G, is in close proximity to the active ₆₁₄KTG₆₁₆ motif.

Although the 15 amoxicillin-resistant isolates shared this unique set of PBP 2B substitutions, they were not clonal, belonging to 5 serotypes (6B, 14, 19A, 19F, and 23F) and 14 PFGE patterns, as well as having variable susceptibility to macrolides, clindamycin, and tetracycline (Table 1).

Most of the PBP changes found in the 25-strain studies were typical of those previously described, such as single or double substitutions in the ₃₃₇STMK₃₄₀ motif of PBP 2X, M₃₃₉F substitution in PBP 2X, and T₃₃₈S or T₃₃₈A substitution in PBP 2X. These substitutions have been described as predominating in strains with intermediate penicillin resistance (2, 5, 15). However, changes in the region 590 to 641 of PBP 2B were unique to the 15 amoxicillin-resistant strains, and transformation experiments have shown that amoxicillin-resistant transformants had acquired multiple PBP changes and that acquisition of these 10 substitutions in PBP 2B by transformants was the key factor associated with amoxicillin resistance (10).

While amoxicillin-resistant strains are currently rare, accounting for only 3.0% of isolates from recent worldwide surveillance studies, and their clinical significance is not clear, the prevalence of these isolates should continue to be monitored. Current formulations of amoxicillin and amoxicillin/clavulanate are active against pneumococci for which amoxicillin MICs are ≤2 μg/ml, and newer formulations, such as amoxicillin/clavulanate suspension (90/6.4 mg/kg of body weight/day) and pharmacokinetically enhanced amoxicillin/clavulanate (2,000/125 mg twice a day) have been developed to address isolates for which amoxicillin MICs are ≤4 μg/ml (7, 11). Further increases in amoxicillin dosing regimens may be required to overcome the resistance described in this report should the prevalence of these strains increase.