TABLE 4.

Ability of AcrA to activate multidrug efflux catalyzed by various E. coli RND transporter^a

Drugb	Length of growth zone (mm)c in an acrA+ (+) or a ΔacrA (−) strain with the following construct:
	pSportI		pAcrB		pAcrD		pYhiV		pMdtBC
	+	−	+	−	+	−	+	−	+	−
Cholic acid	16	17	53	16	36	18	43	19	18	14
Taurocholic acid	23	22	80	19	80	24	80	20	30	25
Novobiocin	0	0d	80	0	20	0	24	0	0	0
Crystal violet	17	23	80	18	20	25	80	22	20	24
Ethidium bromide	26	25	80	24	27	25	80	23	24	22
Ciprofloxacin	19	20	80	16	19	18	80	23	20	20
Erythromycin	45	45	80	40	44	42	80	39	41	40
Tetracycline	8	12	80	14	23	14	16	11	12	10

^aGrowth in the presence of inhibitors was studied by using gradient plates with HNCE1a (acrA⁺) and HNCE1b (ΔacrA::cat). Because chloramphenicol acetyltransferase, encoded by the cat gene, produces resistance to both chloramphenicol and fusidic acid (7), these two antibiotics were removed from the comparison.

^bDrug concentrations used were the same as those in Table 2.

^cSee footnote a of Table 2. Values shown in boldface represent cells in which the presence of AcrA significantly increased resistance.

^dWhen the lengths of growth zones for pSportI-containing HNCE3 and HNCE1b, both acrA disruption or deletion strains, in Table 2 and Table 4 were compared, differences were seen for some drugs, most notably for novobiocin. This result was partly due to the lack of reproducibility in the preparation of gradient plates; comparisons were valid and were made in this study only between strains streaked on the same plate. This result also may have been due to the function of AcrD, which is present in HNCE3 (used in Table 2) but deleted from HNCE1a and HNCE1b (used in Table 4).