act via a salt bridge, thereby neutralizing each other. 
Replacement of either charged residue with a neu- 
tral residue creates an unpaired charge that causes a 
functional defect, while neutral substitutions for 
both residues do not cause inactivation. Conse- 
quently, the secondary-structure model proposed 
for the permease was altered to accommodate a pu- 
tative salt bridge between Asp237 and Lys358 in the 
low dielectric of the membrane, by placing residues 
Phe247 to Thr235 in transmembrane helix VII 
rather than in the hydrophilic domain between he- 
lices VII and VIII. 
As part of an extensive site-directed mutagenesis 
study with an engineered lac permease devoid of 
Cys residues (C-less permease) , putative intramem- 
brane residues Asp237, Asp240, Glu269, Arg302, 
Lys319, His322, Glu 3 2 5, and Lys3 5 8 were systemat- 
ically replaced with Cys; individual replacement of 
any of the residues essentially abolishes active lac- 
tose transport. The single Cys mutants D237C and 
K358C were used to construct a double mutant con- 
taining both Cys substitutions in the same molecule. 
D237C/K358C transports lactose at about half the 
rate of C-less permease to almost the same steady- 
state level of accumulation. Moreover, replacement 
of Asp237 and Lys358, respectively, with Ala and 
Cys or Cys and Ala, or even interchanging Asp237 
with Lys358, causes little change in activity. These 
observations provide a strong indication that 
Asp237 and Lys358 interact to form a salt bridge and 
that neither residue nor the salt bridge is important 
for transport. Despite the relatively high activity of 
the charge-inversion mutant (D237K/K358D) and 
the mutants with neutral substitutions at positions 
237 and 358, immunoblots reveal low levels of the 
polypeptides in the membrane, suggesting a role for 
the salt bridge in permease folding and/or stability. 
The observations also suggest that Asp237 and 
Lys358 may interact in a folding intermediate but 
not in the mature molecule. Remarkably, however, 
an inactive mutant with Cys in place of Asp237 re- 
gains full activity upon carboxymethylation, which 
restores a negative charge at position 237. There- 
fore it seems likely that the interaction between 
Asp237 and Lys358 is important for folding and/or 
stability and that the residues maintain proximity in 
the mature permease. 
To test the possibility that other charged residues 
in transmembrane helices are neutralized by charge- 
pairing, Dr. Kaback and his co-workers constructed 
1 3 additional double mutants in which all possible 
interhelical combinations of negative and positively 
charged residues were replaced pairwise with Cys. 
Only the double mutant D240C/K3 1 9C exhibits sig- 
nificant transport activity. However, the interaction 
between Asp240 and Lys319 is different phenome- 
nologically from that of Asp237-Lys358. Thus 
D240C/K319C catalyzes lactose transport at about 
half the rate of C-less permease to a steady-state level 
that is only ~ 25-30% of the control. Moreover, 
although significant activity is observed with the 
double-Ala mutant or with the two possible Ala-Cys 
combinations, interchanging Asp240 and Lys319 
completely abolishes active transport. Therefore, al- 
though neither Asp240 nor Lys3 1 9 per se is manda- 
tory for active transport, the polarity of the interac- 
tion is apparently important for full activity. Finally, 
unlike the double mutants in D237 and K358, all of 
the D240/K3 1 9 mutants are found in the membrane 
in amounts comparable to the C-less control. 
The results suggest that charge-pairing between 
intramembrane charged residues is probably not a 
general feature of lac permease and may be exclu- 
sive to D237/K358 and D240/K319. However, the 
charge-pair neutralization approach depends on 
permease activity and will not reveal charge-paired 
residues if they are essential for activity. Double-Cys 
mutants involving residues suggested to be H- 
bonded and directly involved in lactose-coupled H"*" 
translocation and/or substrate recognition (i.e., 
Arg302, His322, and Glu325, as well as Glu269, 
which has been shown to be essential) are inactive 
with respect to active lactose transport. 
The modified secondary-structure model of lac 
permease is based on the functional interaction be- 
tween Asp237 and Lys358 and on the notion that the 
intramembrane charged residues are balanced. De- 
spite the indication that Asp240 and Lys319 may 
also participate in a salt bridge, the evidence for the 
interactions is indirect; other approaches are re- 
quired to determine the location of the residues rela- 
tive to the plane of the membrane and to demon- 
strate directly that the pairs are in close proximity. 
C-less permease mutants containing double-Cys or 
paired Cys-Ala replacements will be particularly 
useful. Preliminary efforts to estimate the accessibil- 
ity of Cys residues at positions 237 and 358 with 
water- or lipid-soluble sulfhydryl reagents suggest 
that Cys residues at the two positions are accessible 
to both types of reagents, although the lipid-soluble 
reagents are more effective. Other experiments with 
permease mutants specifically labeled with para- 
magnetic or fluorescent probes, as well as a series of 
alkaline phosphatase fusions in helix VII, are consis- 
tent with placement of both residues near the inter- 
face at the external surface of the membrane rather 
than in the middle of helix VII. Efforts to demon- 
strate disulfide bond formation directly by oxida- 
tion of appropriate double-Cys mutants are also in 
progress. Based on the current evidence, it is clear 
76 
