INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The maltose transport system of the Gram-negative bacterium Escherichia coli is responsible for the unidirectional uptake of maltooligosaccharides ([1,4]-linked D-glucose polymers) (1). This system comprises five proteins. The LamB protein (also known as the maltoporin or -receptor) in the outer membrane, the periplasmic maltose-binding protein (MBP),¹ and the MalF, MalG, and MalK polypeptides that together (MalFGK₂) form the inner membrane complex (2). The LamB protein, encoded by the lamB gene is responsible for the diffusion of maltooligosaccharides into the periplasm of the cell (3). MBP, a soluble periplasmic protein encoded by the malE gene, binds maltooligosaccharides with a high affinity (K_d = 0.1-1 µM) (4, 5) due primarily to a slow dissociation rate (6). MBP is absolutely required for the transport of maltose (7, 8). The three-dimensional structure of MBP has been determined in both the liganded and unliganded forms (9). The MalF and MalG proteins are integral membrane proteins that traverse the membrane eight and six times, respectively (10-12). MalK is an ATPase located on the cytoplasmic side of the inner membrane that is thought to be the energy-coupling component of the maltose transport system (3, 14). MalK shares strong sequence homology with other ATP-binding cassette transporter proteins (16). The stoichiomtry of the inner membrane complex is (MalFGK₂) (17).

MBP has long been regarded as the prime determinant of substrate specificity for the maltose system. However, several observations suggest that the inner membrane complex plays an important role in determining substrate specificity. First, MBP-independent mutants of MalF and MalG have been isolated that transport maltose in the absence of MBP (8). Even in the absence of MBP, these mutants still retain substrate specificity for maltooligosaccharides, which suggests that the inner membrane complex alone is capable of maintaining the specificity of the system. These MBP-independent mutants share a similar K_m for maltose transport of approximately 2 mM, but they each display a V_max that varies from wild-type values to approximately 20-fold lower. Therefore, they all bind maltose equally well, which is suggestive of a similar substrate binding site, but they translocate with different efficiencies.

The observation that the ATP-binding cassette components of two members of the ABC superfamily, the Ugp and Mal transport systems, could be exchanged without any loss of substrate specificity suggests that this component does not play a role in determining the overall specificity of the system (21). The Ugp transport system of E. coli transports sn-glycerol-3-phosphate. The UgpC and MalK proteins of these systems are highly homologous and both couple energy via ATP-hydrolysis (21). Thus, the two other components of the maltose transport system inner membrane complex, MalF and MalG, must be contributing to the specificity of the system.

Mutants of MalF have been isolated that alter the range of substrates transported by the maltose transport system (22). The various mutations map to the malF gene and cause alterations in transmembrane domains 6, 7, and 8 of MalF. The mutants could only recognize either maltose or longer maltodextrins, but not both. The mutations cluster along the transmembrane helices, and suppressor mutations in neighboring helices of MalF suggest a physical interaction.

Studies on other members of the ABC superfamily, such as P-glycoprotein and the HlyB transporter, have also suggested that the transmembrane helices play an important role in determining substrate specificity (23-27).

In the present study, we describe a mutant of the maltose transport system (MalF540) that transports lactose efficiently. The MalF540 mutant we isolated carries a mutation in the malF gene that changes a glutamine codon at position 99 of MalF to an amber stop codon. This mutant suggests a novel mechanism for altering substrate specificity. The isolation and characterization of this mutant with respect to transport efficiency, maltose transport components required, and substrate specificity is described.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Media and Genetic Techniques-- Rich media (LB) and minimal media (M63) were prepared as described previously (30). Standard genetic procedures were performed by the method of Miller (30) unless otherwise noted. Maltose and lactose were obtained from Pfanstiehl Laboratories, Inc. All other sugars were obtained from Sigma. The following antibiotics were used at the indicated concentration unless otherwise noted: 100 µg/ml carbenicillin, 100 µg/ml ampicillin, 50 µg/ml kanamycin, 25 µg/ml chloramphenicol, 20 µg/ml tetracycline.

Bacterial Strains-- Bacterial strains are listed in Table I. The recombination defective recA1 strain, GM1418, was constructed by P1vir transduction using a lysate prepared from HS3078 containing the srl::Tn10 mutation closely linked to the recA1 allele. Tetracycline-resistant transductants of LH1375 were selected and screened for UV sensitivity.

malF argE

malF3

malF3

Plasmid Construction-- The plasmid pGM1, which carries the malE gene under control of the ptac promoter, was constructed by ligating the malE-containing EcoRI/StuI fragment from NF630 (31) phage DNA, to EcoRI/SmaI-digested pMMB207 plasmid vector DNA. pGM7, which carries the malK and malG genes under the malB promoter, was constructed by ligating the malK-containing EcoRI/ScaI fragment from pHS4, to the malG-containing EcoRI/SnaBI fragment from pMR24. pLac2SM was constructed by digesting pLac2 with SacI and MfeI, thus deleting the portion of malF distal to the specificity mutation, "end-filling" and "chewing back" the respective overhangs with DNA polymerase I (Klenow) and T4 DNA-polymerase, and then ligating the resulting large DNA fragment to itself. pLH9SM was constructed in the same way except that the pLH9 plasmid was used. pLac2F was constructed by ligating the PstI/BsmI fragment from pLac2 containing the 5-end of the truncated malF540 gene including the mutation, to the PstI/BsmI fragment from pBR322 that contains the origin of replication. Pt-pLac2SM was constructed by ligating the PstI/BstEII fragment from pMR38 containing the Ptrc promoter and the 5-end of the malF gene to the PstI/BstEII fragment from pLac2SM containing the 3-end of truncated malF540 and malG.

Isolation of the Lactose Specificity Mutant-- A plasmid carrying the malF502 MBP-independent allele, pLH5, was mutagenized in the mutator strain KD1087 (32). This strain contains the mutD5 allele, which increases the frequency of mutational events 50-100 times above that observed in a wild-type strain (mut⁺). KD1087 was transformed with the pLH5 plasmid, and transformants were selected on LB plates containing carbenicillin. Individual transformants were then grown in LB containing carbenicillin at 37 °C overnight. Plasmid DNA was purified from these independent cultures and used to transform the lactose indicator strain GM1305. Transformants were plated on minimal 1% lactose plates containing carbenicillin, kanamycin, chloramphenicol, and 0.2% arginine. Those transformants that grew on the indicator media after 1-3 days of incubation at 37 °C were purified by passing three times on minimal 1% lactose plates. Plasmid DNA was purified from these isolates and used to retransform the lactose indicator strain to ensure that the mutation responsible for the Lac⁺ phenotype is linked to the pLH5 plasmid.

Nucleotide Sequence Determination-- The nucleotide sequence of malF and malG alleles carried on pLac2 and various amber mutant plasmids (pGM8-pGM12) was determined by using double-stranded DNA templates and 13 oligonucleotide primers (33). Double-stranded DNA templates were purified, denatured, and annealed with the primers as described (34). The Sanger dideoxynucleotide chain termination method (35) was used to determine the nucleotide sequences.

Transport Assays-- Maltose and lactose transport activity was estimated by measuring the uptake of [¹⁴C]maltose (360 mCi/mmol) or [¹⁴C]lactose (57 mCi/mmol) as described previously (7). When sugars were tested for their ability to inhibit radioactive sugar uptake, cells were incubated in the presence of inhibitor sugar for 5 s prior to the addition of radioactive substrate. [¹⁴C]Maltose was obtained from Moravek Biochemicals, Inc. [¹⁴C]Lactose was obtained from Amersham International, plc.

Assay of ATPase Activity in Inside-out Membrane Vesicles-- To measure the ATPase activity of the mutant maltose systems, the respective proteins must be overproduced. We used the strain HS3309, which carries the pMR11 plasmid that has the malK gene under the Ptac promoter. This strain was transformed with either pBR322 as a vector control; pNT11, which carries the malF502 MBP-independent allele under Ptrc; or pt-Lac2SM, which carries the malF540 allele under Ptrc. Inside-out membrane vesicles were prepared from transformants induced with isopropyl-1-thio--D-galactopyranoside, and ATPase activity was measured as described (14, 36). Protein concentration was determined using the BCA protein assay reagent kit from Pierce.

Construction of the Various malF-amber Mutants-- The plasmids carrying mutations in malF resulting in amber stop codons in place of amino acids Glu³⁹, Tyr⁵⁵, Glu¹³⁰, Lys²⁷⁵, and Tyr³¹⁷ were constructed by site-directed in vitro mutagenesis (37). The following oligomers were used in this experiment: GM1, 5-GGCGAACAGGTACTACCCTTGTGCG-3, Glu³⁹ to amber stop; GM2, 5-CGATTGGCGAAAATCTACAGCCCCG-3, Tyr⁵⁵ to amber stop; GM3, 5-CGCCAGTTGCCACTAATCGCCCGC-3, Glu¹³⁰ to amber stop; GM4, 5-GGCGAGGAACGGCTACTGAATGCC-3, Lys²⁷⁵ to amber stop; GM5, 5-GCAGGACGCTAGACCGGCTTTGC-3, Tyr³¹⁷ to amber stop.

Protein Gel Electrophoresis-- Gel electrophoresis was carried out by the method of Laemmli (39). Samples were diluted in sample buffer containing 5% (v/v) 2-mercaptoethanol and heated in a boiling water bath for 5 min. Electrophoresis was carried out in 12% polyacrylamide gels. Protein was visualized by staining the gels with 0.2% Coomassie Brilliant Blue R-250 in methanol-acetic acid-water (5:1:5) and destaining in 7.5% acetic acid, 5% methanol.

Membrane Protein Solubilization-- We examined the ability of Triton X-100 to solubilize the MalFGK proteins as described previously (14) with some alterations. 250 µg of protein from the inside-out membrane vesicles was incubated in 1.0 ml of 50 mM Tris-HCl, pH 7.5, 1 mM dithiothreitol, 5 mM MgCl containing 2% (v/v) Triton X-100 for 1/2 h. The samples were centrifuged at 14,000 × g in a microcentrifuge for 20 min. The supernatant was carefully removed, and the pellet was resuspended in 50 µl of 100 mM Tris-HCl; 5 µl of this sample was loaded on gels for Coomassie staining and 15 µl for Western blots after boiling for 4 min along with 3 µl of 50% glycerol and 10 µl of sample buffer. The supernatant was added to 2.0 ml of cold ethanol and placed on dry ice for 1 h. The samples were centrifuged at 12,000 × g in a microcentrifuge for 30 min at 4 °C. The ethanol supernatant was removed, and the percipitated proteins were resuspended in 40 µl of 100 mM Tris-HCl, pH 7.5. 6 µl of this sample was loaded on gels for Coomassie staining and 17 µl for Western blots after boiling for 4 min along with 3 µl of 50% glycerol and 10 µl of sample buffer.

Immunodetection by Western Blotting-- Proteins were transferred from SDS-polyacrylamide gels to sheets of nitrocellulose (BAS 0.45-µm pore size, Schleicher & Schuell) by electroblotting (20 V, overnight) as described elsewhere (40). The nitrocellulose sheets were blocked with 5% (w/v) nonfat powdered milk in TBST for 1 h and then incubated with the appropriate dilution of anti-MalK or anti-MalF rabbit antibody. After extensive washing with TBST, goat anti-rabbit IgG coupled to peroxidase was added for 1 h. The resulting blot was then developed using the ECL-Western blotting kit from Amersham Life Science.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Isolation of the Lactose Specificity Mutant-- The indicator strains used to detect altered substrate specificity mutants of the maltose transport system that transport lactose share some common attributes. They all have a deletion of the chromosomal lactose operon but carry the lacZ gene on a plasmid or F-episome. Thus, these strains can metabolize intracellular lactose but are incapable of transporting it into the cell. The strains also carry the malTp1Tp7 allele (41) ensuring constitutive mal gene expression in all growth media.

Mapping of the Lactose Specificity Mutant-- The mutation responsible for the phenotype of the MalF540 mutant was mapped by performing a series of DNA fragment exchanges between the pLac2 plasmid, which carries the malF540 allele, and the pLH5 parent plasmid (Fig. 1). These exchanges suggest that the mutation is located proximal to the SacI site on the pLac2 plasmid, somewhere at the very beginning of malF. Not shown in Fig. 1 is the second restriction site used in these experiments, a PstI site located in the ampicillin resistance gene.

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Mapping of the lactose specificity mutation. Fragment exchanges between pLac2 and the parent MBP-independent plasmid pLH5 or the wild-type plasmid pLH9 are shown. The lactose phenotype was assayed as growth on minimal 1% lactose plates containing ampicillin after 2 days at 37 °C in strain GM1305.

View larger version (46K): [in this window] [in a new window]   Fig. 2.   Topological map of the MalF and MalG proteins. This model is based on the results obtained with malF-phoA and malG-phoA hybrids. The positions of the amino acid residues altered in the lactose specificity mutant MalF540 (Q99Amber) and the MBP-independent mutant MalF502 (G338R and A502V) are indicated.

The Lactose Specificity Mutant Does Not Transport Maltose-- To find out if the mutation had any effect on the ability of cells to utilize maltose as a sole carbon and energy source, we evaluated the growth of a strain that makes wild-type MBP (GM1035) and a strain that does not (GM1042) on minimal maltose plates when transformed with pLac2 (Table III). These strains carry a chromosomal Tn5 insertion in malE that is polar on malF and malG but does not affect the intact chromosomal malK gene. A plasmid that carries the wild-type malF and malG genes (pLH9) conferred the ability to grow on maltose only in the strain background that contains MBP (GM1035). This was expected, since MBP is absolutely required for the wild-type transport of maltose. The pLH5 plasmid that carries the malF502 MBP-independent allele only permitted growth in the strain background that does not contain MBP (GM1042). MBP-independent mutants, including MalF502, were selected on the basis of being able to utilize maltose in the absence of MBP. MBP-independent mutants cannot utilize maltose in the presence of MBP at physiological concentrations due to a faulty interaction of MBP with the inner membrane complex (MalFGK₂) (8). The pLac2 plasmid did not confer a Mal⁺ phenotype to either strain. It has thus lost the MBP-independent phenotype associated with its parent plasmid pLH5.

Lactose Transport by the Lactose Specificity Mutant Requires MalK-ATPase Activity but Does Not Require MBP-- We compared the lactose phenotype of MalF540 in both wild-type (GM1265) and malB101 deletion (GM1191) strain backgrounds carrying the pLH22 plasmid that provides MalK. The malB101 deletion covers all of the genes required for the transport of maltose across the cytoplasmic membrane. The plasmid carrying the malF540 allele and malG⁺, pLac2, conferred the ability to utilize lactose as a sole carbon source in both the GM1265 and GM1191 strains, as assayed by growth on minimal lactose plates (Table II). Since in both strains MalG and MalK are being provided from either a plasmid or from the chromosome, but only GM1265 produces MBP, this result suggests that MBP is not required for a Lac⁺ phenotype.

The Lactose Specificity Mutant Does Not Require Full-length MalF to Transport Lactose-- To determine if the region of malF distal to the amber stop codon of MalF540 is required and to evaluate the possibility of a restart protein playing a role in the mutant phenotype, the remaining portion of malF was deleted from the plasmid pLac2, and the resulting plasmid, pLac2SM, was assayed for a lactose phenotype on minimal lactose plates (Table II). The region deleted was from a SacI site located 300 base pairs distal to the amber stop codon, to an MfeI site at the very end of malF. The pLac2SM plasmid still displays a Lac⁺ phenotype comparable with the pLac2 plasmid when introduced into the lactose indicator strains GM1305 and GM1191. The fact that the pLac2SM plasmid displays a Lac⁺ phenotype in the deletion strain GM1191 indicates that full-length MalF is not required in conjunction with the amber mutant to transport lactose. To ensure that the deletion itself does not cause a Lac⁺ phenotype, the same deletion was performed on the wild-type plasmid pLH9. The resulting plasmid, pLH9SM, did not display a Lac⁺ phenotype.

The Lactose Specificity Mutant Requires MalG to Transport Lactose-- To ascertain if MalG is required for the transport of lactose by MalF540, we took a small fragment containing the malB promoter and the beginning of malF540 carrying the Q99(Am) mutation from pLac2 (up to the BsmA1 site, which is 300 base pairs distal to the mutation) and cloned it into pBR322. The resulting plasmid, pLac2F, did not display a Lac⁺ phenotype when introduced into the strain GM1191, which is deleted for the chromosomal malB region and has MalK provided off of a plasmid (Table II). However, the plasmid pLac2SM, which carries a similar portion of malF540 and malG⁺, displays a Lac⁺ phenotype in the same strain, suggesting that MalG is required for the Lac⁺ phenotype observed with the pLac2 plasmid.

Growth and Transport Properties of the Lactose Specificity Mutant-- To measure the ability of MalF540 to utilize lactose as a sole carbon and energy source, we determined the generation time of a strain carrying the pLac2 plasmid in M63 minimal lactose liquid media (Table IV). We also measured the ability of the mutant to transport [¹⁴C]lactose (Table IV).

The Specificity of Transport by the Lactose Specificity Mutant-- We examined the specificity of lactose transport by MalF540 by evaluating the ability of different sugars to compete with [¹⁴C]lactose uptake (Fig. 3). A collection of 30 available monosaccharides, disaccharides, and trisaccharides was tested at 100 mM to see if any of the sugars inhibited the transport of 1 mM [¹⁴C]lactose (data not shown). Sugars that displayed the greatest inhibitory activity were then tested at 10 mM inhibitor concentration (Fig. 3). Only a small number of sugars strongly inhibited [¹⁴C]lactose transport at the lower concentration. Those sugars that were tested for inhibitory activity at 100 mM but did not noticeably inhibit included methyl glucopyranoside, arabinose, rhamnose, fructose, xylose, melibiose, raffinose, palatinose, cellobiose, gentiobiose, isomaltose, and fucose.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Inhibition of [14C]lactose uptake by various sugars. Uptake of [14C]lactose by whole cells of the lactose specificity mutant MalF540 was measured at a [14C]lactose concentration of 0.5 mM in the GM1305 background. The concentration of competitor sugar was 10 mM.

The Lactose Specificity Mutant Displays Constitutive ATPase Activity-- We were interested in characterizing the ATPase activity of the MalF540 mutant, because it is thought that MBP is responsible for triggering the ATPase activity of MalK and subsequent transport by the inner membrane complex (18). Since the MalF540 mutant does not require MBP for lactose transport, we speculated that this mutant might display constitutive ATPase activity in analogous fashion to the MBP-independent mutants described earlier.

View larger version (43K): [in this window] [in a new window]   Fig. 4.   The presence of MalK in whole cells used to make everted membrane vesicles. Equal amounts of whole cells of the strain CHP243 transformed with the indicated plasmid were subjected to electrophoresis on polyacrylamide gels containing SDS. Lanes 1 and 4, PBR322 control; lanes 2 and 5, pNT11 (malF500 MBP-independent mutant allele); lanes 3 and 6, ptac-Lac2s (malF540 lactose specificity mutant allele (Q99Amber)); lanes 4, 5, and 6, from a Western blot probed with anti-MalK antibodies.

View larger version (33K): [in this window] [in a new window]   Fig. 5.   Presence of MalK in detergent-soluble fraction of everted membrane vesicles. Equal amounts, with respect to total protein concentration, of everted membrane vesicles of the strain CHP243 transformed with the indicated plasmid were solubilized with Triton X-100. Detergent-soluble fractions were subjected to electrophoresis on polyacrylamide gels containing SDS. Lanes 1, 4, and 7, pBR322 control; lanes 2, 5, and 8, pNT11 (malF500 MBP-independent mutant allele); lanes 3, 6, and 9, ptac-Lac2s (malF540 lactose specificity mutant allele (Q99Amber)). Lanes 4, 5, and 6 are from a Western blot probed with anti-MalF antibodies. Lanes 7, 8, and 9 are from a Western blot probed with anti-MalK antibodies.

Other Amber Mutants and Their Ability to Transport Lactose-- To map the extent of MalF protein that is necessary for promoting lactose transport, we constructed five amber mutants by performing site-directed mutagenesis on the plasmid pLH9, which carries the wild-type malF and malG genes (Fig. 6). E39(Am) contains an amber stop codon after the first transmembrane helix of MalF. Y55(Am) contains an amber stop codon after the second transmembrane helix of MalF. Q99(Am) contains an amber stop codon after the third transmembrane helix of MalF, and this is the original mutation responsible for the Lac⁺ phenotype conferred by the malF540 allele. E130(Am) contains an amber stop codon in the large periplasmic loop, after the third transmembrane helix of MalF. K275(Am) contains an amber stop codon after the large periplasmic loop of MalF. Y317(Am) contains an amber stop codon after the fourth transmembrane helix of MalF. We examined the maltose and lactose phenotypes of these amber mutants and measured their ability to transport [¹⁴C]lactose.

View larger version (34K): [in this window] [in a new window]   Fig. 6.   Transport of [14C]lactose by various amber mutants in MalF. Uptake by whole cells was measured at a [14C]lactose concentration of 2 mM in the GM1305 strain background.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

We have isolated a mutant of the maltose transport system that alters its substrate specificity. The wild-type system is incapable of transporting lactose, but the mutant transports lactose and has lost the ability to transport maltose. A Q99(Am) mutation in the MalF protein is responsible for the Lac⁺ phenotype. In addition, two sugars that do not inhibit transport by the wild-type or MBP-independent transporters do inhibit lactose transport by the MalF540 complex (4, 8).

The fact that the lactose specificity mutant does not require MBP and the first three transmembrane helices of MalF, MalG, and MalK are sufficient to transport lactose, while still retaining substrate specificity, suggests that the integral membrane components of the mutant alone are sufficient to form a substrate site and determine substrate specificity. Recent experiments on the P-glycoprotein (P-gp) system, a eukaryotic ABC transporter, have provided additional evidence that the integral membrane components play an important role in determining the substrate specificity of the system. These experiments include the genetic analysis of chimeric molecules constructed from different members of the P-gp family (43), the isolation and construction of discrete mutations in P-gp (27), epitope mapping and cross-linking studies with substrates and substrate analogs of P-gp (24, 25, 44), as well as energy transfer experiments (23).

[¹⁴C]Lactose transport experiments with a collection of amber mutants at strategic locations along MalF indicate that the first transmembrane segment of MalF along with MalG (malF541; E39(Am)) and MalK is sufficient to transport lactose (~67% of malF540) (Fig. 6), while MalG and MalK alone cannot transport lactose (data not shown). Solubilization studies have indicated that MalF and MalK can form a stable complex in the inner membrane but that MalG and MalK do not (14). Others have shown that the first transmembrane helix of MalF can be deleted without seriously affecting the transport of [¹⁴C]maltose (~60% of wild type) (45). Since the first helix is not essential for maltose transport, yet the lactose specificity mutant requires this helix for lactose uptake, it is highly unlikely that this helix is directly involved in the transport of substrate. Rather, it is more likely that the helix is allowing MalG to assume the correct conformation in the membrane that is solely responsible for the recognition and transport of lactose. Very similar results have been observed with LacY. A large portion (N-terminal 22 amino acids) of the first transmembrane helix of LacY was found not to be obligatory for lactose transport activity (46). However, a deletion construct of LacY, which contains only the first N-terminal transmembrane helix and the last six C-terminal transmembrane helices, was found to transport lactose efficiently (~80% of wild type) and specifically (47). Furthermore, with over 95% of the residues of LacY mutagenized, the four charged residue that have been shown to be mandatory for lactose transport are located in transmembrane domains of the C-terminal half of the protein (48). Much as in our lactose specificity mutant, a pathway for lactose transport in LacY can be formed solely by the last six putative transmembrane helices.

A mutant of another member of the ABC transporter superfamily, the cystic fibrosis transmembrane conductance regulator, only contains the first six transmembrane helices, the first nucleotide binding domain, and the regulatory domain, yet it is still able to form a regulated ion channel in HeLa cells (49). Sucrose gradient sedimentation studies indicate that the "half molecule" truncation mutant forms homodimers. This mutant is similar in structure to our lactose specificity mutants, in that it only has half of the wild-type transport apparatus, and suggests a possible model for its activity. A subunit composed of the MalG protein and the stabilizing small portion of MalF may form a homodimer with another such subunit, thus forming a transport path for substrate.

The results with the MalF amber mutants indicate that the ability to transport lactose is compromised once a portion of the periplasmic loop between the third and fourth transmembrane helices is included in the construct (Fig. 6). Inclusion of this loop might block the access of the transport path by lactose, or the loop might simply not allow the MalG protein to assume the correct conformation in the membrane due to a steric constraint. Alternatively, the loop might compromise the constitutive ATPase activity of the MalK subunit. Although the periplasmic loop has been shown to be essential for the transport of maltose,² no specific function has ever been assigned to this domain of the transport apparatus.