Determine the cadmium resistance of cells harboring pKJ100

In order to determine the cadmium resistance of those cells harboring plasmid pKJ100, in which the full-length cadA gene was placed under control by a trc promoter, cells were grown in the presence of different concentrations of cadmium. The survivals of these cells were measured their optical densities at a wavelength of 600 nm

spectrophotometrically after a 6-hr of incubations. As shown in the Fig. 6, cells contain plasmid pKJ100 grew well in the presence of higher

concentration of Cd²⁺; however those cells without the pKJ100 plasmids display cadmium sensitivities (at least 10 times less in resistance).

Suggesting that CadA is functionally expressed in this system and

implying the fact that the CadA protein is naturally folded in bacterial cell membrane.

3.2 Computer hydropathy analysis

In this thesis study, our major goal is to investigate the CadA

topography. Using the method of Kyte and Doolittle (1982) and computer program of TMpred (Hofmann and Stoffel, 1993), a predicted CadA topography was drawn based on the hydrophathy profile of the 727 amino acid sequences of CadA (Fig. 2a, 2b). And there are six transmembrane segments (TMs) in CadA protein were identified and a CadA protein model was proposed (Silver and Walderhaung 1992). In this postulated

37

Fig. 6. Cadmium resistance assays and SDS-PAGE analysis of wild-type CadA in pKJ100. (A) Cadmium resistance assays were performed using E. coli strain RW3110

harboring pKJ100, wild-type cadA gene, or the parental plasmid, pSE380, to grow in the presence of different concentrations of cadmium as described. The survival rates were calculated by dividing the OD600 of the final reading in culture with each given cadmium concentration by the OD600 measured from culture without the cadmium addition, and shown as percentage (%) in the figure. The average absorbances of the cultures without cadmium addition were 1.82 and 1.94, respectively. The SD values in these results were all less than 8%. (B) CadA protein was visualized in SDS 7%-polyacrylamide gel electrophoresis by comparing the membrane proteins prepared from cells harboring pKJ100 and pSE380 plasmids. M, M.W. markers; 1, cells containing pSE380 plasmid; 2, cells harboring pKJ100 plasmid.

%

CadA topography, 6 highly hydrophobic stretches of amino acid residues of the highly possible membrane-spanning domains including amino acid sequences at 106-126, 130-150, 336-356, 364-384, 626-646 and 686-706 were suggested (Silver and Walderhaung 1992). However, from

hydropathy profile of CadA, there might be another 3 TMs, which are located at the amino acid positioned 160-179, 180-196 and that of amended from the latest domain, 677-697 and 699-719, should be included in the CadA topography. The possibility of the presence of additional TMs in CadA was examined in this study. Experiments were designed to prepare CadA fusions with the reporter enzymes which was placed in each possible loop structure and facing either the periplasm or the cytoplasm location. Furthermore, additional CadA fusions were prepared in the putative transmembrane domains and in the large cytoplasmic portions of the protein in order to clarify the CadA topography.

3.3 Isolation of cadA-phoA and cadA-lacZ fusions

To study the membrane topography of CadA, phoA and lacZ genes were used as enzyme reporters and were separately fused at 22 different sites on CadA using PCR cloning method as described in the

“Experimental Procedures”. The 44 cadA fusion plasmids were transformed in E. coli JM109 cells and the successful clones were selected in agar plates in the presence of BCIP or X-gal. The phoA fusions at the CadA amino acid positions 130, 181, 356 and 703 were isolated from the blue colonies on the BCIP plates. Suggesting that these clones contain the translational alkaline phosphatase (AP) on their

39

periplasmic locations of the cells. Other cadA-phoA fusions were isolated from those pale colonies and confirmed by DNA sequence determinations, which indicating that the AP in these fusions were located within the

cytoplasmic side in their harboring cells (data not shown).

To isolate the cadA-lacZ fusions, we grew cells in the agar plates with X-gal. Most of the successful clones display a blue-color colony formation and were picked up for DNA sequence determination to confirm the fusions. These fusions implied their cytoplasmic location of the CadA junction region with the LacZ moieties. However, fusions at the CadA positions 130, 356, 388 and 703 produced pale colonies, suggesting that these later clones contain periplasmic locations of LacZ moieties of the fusions (data not shown).

The nucleotide sequences of the cadA-phoA and cadA-lacZ fusion junctions are listed in Table II.

3.4 Determination of the first transmembrane segment

The N-terminal first 130 amino acids were determined their locations using immunoblotting procedure. Two plasmids, pL109P and pL130P, in which a PhoA reporter protein was fused at the amino acid positions 109 and 130 respectively, were prepared and were transformed into E. coli strain LMG194. Cells harboring either pL109P or pL130P plasmid were grown and their cell components were fractionated. Using

immunoblotting method, only those either cytosolic fractions prepared from cells containing pL109P or membrane fraction prepared from cells with pL130P plasmid were found immunoreactive band when probed with anti- E. coli alkaline phosphatase antibody. As shown in the Fig. 7,

Table III Nucleotide sequence of the cadA-phoA and cadA-lacZ fusion junction.

The fusion site, BamHI, is underlined. Left of the fusion sites are the sequences of cadA PCR fragments. Right of them are ether phoA or lacZ sequences. The lowercase letters represent altered nucleotides on cadA.

Plasmid CadA-phoA fusion junction pL109P

Plasmid CadA-lacZ fusion junction pL109L

41

Fig. 7. Determination of the first transmembrane segment of CadA by western blotting.

Cultures of E. coli strain LMG194 bearing the first and second cadA-phoA fusion plasmids, pL109P and pL103P, were induced by 0.1 mM IPTG for 2 hr. Cells were fractionated into cytosol and membranes as described. Samples for SDS-PAGE were the whole cell lysates of control cells, and the cytosol and membranes of the fusion plasmid-bore cells. The samples were analyzed on a 7 % polyacrylamide gel followed by western blotting with an antibody to alkaline phosphatase. The hybrid proteins that could be detected on the blot were in the cytosol of pL109P-bore cells and the membranes of the L130P-bore cells

immunoreactive bands with a molecular weight of 63.4 kDa and 65.7 kDa were observed in cytosolic fraction of pL109P clone and membrane

fraction of pL130P clone, respectively. The proximal molecular weights calculated from the blot were close to the predicted hybrid proteins and these data suggested that the N-terminal first 109 amino acids were located within the cytoplasmic portion of the cell, and the first TM

occurred approximately between CadA amino acid sequences 109 to 130.

3.5 Enzyme activity analysis of cadA-phoA and cadA-lacZ fusions

Other than the blue-white selection of the colony formation to determine the locations of the reporter enzymes, we have also performed enzymatic assays to measure the enzyme activities of those fusion

proteins. Two reporter genes, phoA and lacZ, were used for this topology study, and each of the reporter genes was fused at twenty-two different locations throughout the cadA gene to generate either as cadA-phoA or cadA-lacZ fusions. These genes are transformed into E. coli LMG194 cells, for expression of cadA-phoA gene fusions, and E. coli MC1000 cells, for expression of cadA-lacZ gene fusions. Either membrane or cytosolic fraction prepared from cells harboring these fusion genes was assayed for their alkaline phosphatase or β-galactosidase activities. As shown in the Table IV, two classes of enzymatic activities were found among those cells with alkaline phosphatase fusions. Including those fusions of L130, E181, W356, and A703 display a high alkaline

phosphatase activity, and the activities were ranging from 51 to 200 AP activity units (Table IVa). Suggesting that the expressed reporter enzymes from these fusions were located in the periplasmic region of the cells. On

43

Table IV. Fusion enzyme activities. (a.) Alkaline phosphatase activity of E. coli

strain LMG194 expressing cadA-phoA fusions. (b.) β-galactosidase activity of E. coli strain MC1000 harboring cadA-lacZ fusions. *Activity = (△A420‧1000) / (min‧ml‧

A600), 23ºC. Standard deviations (SD) are given between parentheses.

a.

the other hand, eighteen of these fusions, which include L109, L155, R204, V230, V255, A283, V326, G388, F414, I445, V477, N509, V542, Q577, E611, I646, F681, and D726 showed no enzymatic activities at all.

Indicating that the expressed alkaline phosphatase from these fusions were at the cytoplasmic side of the cells.

Totally twenty-two cadA-lacZ fusions at the same positions as those of phoA fusions were prepared in E. coli MC1000 cells in this study. And the expressed β-galactosidase activities from these fusions were also investigated. In these β-galactosidase activity assays, we found that most of those fusions with a low alkaline phosphatase activity in cadA-phoA clones displayed high β-galactosidase activities when fused with lacZ gene, except for the G388 fusion, and their activities were ranging from 12 to 94 units of β-galactosidase activity (Table IVb). As we predicted, those periplasmic located fusions demonstrated above have a low β-galactosidase activity with an exception of the fusion at the E181 position of CadA. The reasonable explanation for the observation is that the expressed β-galactosidases resulted in cadA-lacZ fusion at the G388 position of CadA might be due to the closely location of the enzyme to the membrane and the high β-galactosidase activity found in E181 fusion might be due to its incomplete transmembrane. Other than these two fusions, the β-galactosidase activity assays are in agreement with those found in the PhoA enzyme activity assays (Fig. 8).

3.6 Determine the protein productions of CadA fusion

In order to determine the production of those CadA fusion proteins used in this study, Western blot analysis of those fusion proteins were

45

a.

Fig. 8. Topological structure of the CadA protein. (a.) A model with enzyme activities is shown. Alkaline

phosphatase activity is shown in a dark rectangle, while the β-galactosidase activity is shown in a light rectangle (b.) Shown with conserved motifs belonged to CPx-ATPases, the model indicates these motifs by light-colored circles and ellipses. The locations of the phoA and lacZ fusion sites are indicated as filled circles, and expressed by a one-letter code and the residue number. The fusion sites are fairly distributed on the CadA protein, except the first 109 residues, which were verified else by western blot (Fig. 8). The first 109 residues as well as the amino acids in the 8 certain transmembrane segments are shown by open circles.

Charged amino acids around the transmembrane segments are also marked here.

b.

performed. Either membrane or cytosolic fractions prepared from cells harboring plasmids with either cadA-phoA or cadA-lacZ fusions were grown and induced in the presence of 0.1 mM IPTG as described. For all forty-four fused protein preparations were examined using monoclonal antibody against either alkaline phosphatase or β-galactosidase. As shown in fig. 9, all fusion proteins of both CadA-alkaline phosphatase (Fig. 9a) and CadA-β-galactosidase (Fig. 9b) were detected using the

immunoblotting analysis. The proximal molecular weights calculated from the blot were close to the predicted fusion proteins, and the molecular weights of the fusion proteins are listed in Table V.

47

Fig. 9. Western blots of cell extracts of various fusions. (a.) E. coli strain LMG194 expressing various

cadA-phoA fusions. (b.) E. coli strain MC1000 expressing various cadA-lacZ fusions. Samples for

SDS-PAGE were the whole cell lysates of control cells, the cytosol of pL109P- and pL109L-bore cells, and the membranes of other cloned cells. The solubilized samples were analyzed on 7 % polyacrylamide gels, and immunoblotting with antibodies to alkaline phosphatase and β-galactosidase, respectively.

Table V. Molecular weights of the fusion proteins. (a.) CadA-alkaline

phosphatase fusions. (b.) CadA-β-galactosidase fusions.

a.

49

Chapter 4. Discussion

To understand the structure of membrane transport proteins and their orientation across the membrane is necessary for determining the

molecular mechanism responsible for the phenomenon in a membrane transport system. There are many techniques available for the membrane topology determination, including chemical labeling, immunodetection, in vitro transcription/translation, and reporter gene fusions. In this thesis study, the structure of the CadA protein was examined experimentally using the genetic approach to construct various cadA gene fusions with the genes encoding either alkaline phosphatase (phoA) or β-galactosidase (lacZ) enzymes at different locations of the protein. This approach of using gene fusions to elucidate the architecture of membrane proteins has been applied in many protein topographical studies (Manoil, 1988 and 1991; Lee & Manoil, 1996).

Two reporter gene fusions were used in this study separately. The results of the two types of fusions were all consistent to each other and were used to propose the topological model as shown in Fig. 7. In this model, 8 transmembrane segments (TMs) were identified, along with 3 cytoplasmic loops, and 4 periplasmic loops were also found. In this model, the N-terminal domain of approximately 100 amino acid residues and the C-terminal domain of CadA protein were shown to be located in the cytoplasmic side. Taken together with the hydropathy data (using SOAP and TMpred method to predict) and positive-inside rule (von Heijne, 1992), we proposed that the TMs of CadA protein should be amended at the amino acid positions ranging from 105-123 (TM1),

131-151 (TM2), 164-192 (TM3+TM4), 332-356 (TM5), 363-391 (TM6),

677-697 (TM7), and 699-719 (TM8).

On the other hand, our data have demonstrated a difference in TM3 and TM4 when compared it to that of the Helicobacter CadA (Melchers et al., 1996, 1999). For example, the amino acid sequence of 164-192, including about 30 amino acids was originally predicted to be as a low hydrophobic domain, and seems unable to traverse the lipid bilayer by separating into two independent TMs (Fig. 2). Our data have shown that fusion enzyme activities on the fusion site E181, at the middle of the region, were asymmetric in this study. As we have predicted that enzyme activity of phoA fusions at E181, pE181P fusion, showed a high alkaline phosphatase activity (Table IVa), suggesting that the amino acid is at the periplasmic side as that of Helicobacter CadA (Melchers et al., 1999).

However, using lacZ fusion at the position, the pE181L clone, exhibited an unexpected high β-galactosidase activity which result disagrees with the data from phoA fusion at the same position and the Helicobacter CadA (Melchers et al., 1999; Table IV). The difference could be

explained that the periplasm-tended alkaline phosphatase is much easily to translocate the fusion site, E181, out to the periplasm; however, the larger cytoplasm-tended β-galactosidase would be kept in its cytoplasmic location due to the short hydrophobic region between the amino acid positions of 164 and 181 can not make the enzyme translocation.

Therefore, we believe that the region including the amino acid sequences from 164 to 192 are associated with the cytoplasmic membrane, but not traverses across the membrane. There is another possible explanation that the region is arranged around the hydrophilic channel of the transporter,

so it can be separated into two TMs in spite of its low hydrophobicity.

In general, our CadA topological results are similar to that of the

51

previously reported Helicobacter pylori homologous (Melchers et al., 1996, 1999). In these two CadA models, we found that both CadAs include 8 TMs and spanning on the membrane in a similar distribution of those common motifs (Fig. 8b). The conserved phosphatase domain (TGES motif) special to P-type ATPases is located between TM4 and TM5, while the signatures of phosphorylation (DKTGT motif) and ATP binding (GDGXNDXP motif) are positioned between TM6 and TM7. The conserved CPC motif found in all CPx-type ATPases is located within the TM6, as it was previously predicted (Solioz and Vulpe, 1996) and has demonstrated by the Helicobacterial CadA (Melchers et al., 1996, 1999).

On the other hand, there are some differences between these two CadAs. The staphylococcal CadA differs from the Helicobacter CadA in one of its highly hydrophobic regions between amino acid positions 626 and 646 in Staphylococcal CadA, but not found in its Helicobacter counterpart. This region was predicted as a putative TM by computer analysis but has not found in Helicobacter CadA (Fig. 2). Although our data did not support the presence of this putative TM, this hydrophobic region, on the other hand, might associate to the membrane and

participate the CadA function. Another difference is the high hydrophilic region found in the N-terminal of CadA protein, just before the TM1 and between amino acid positions 81 and 102. Within this region, the

staphylococcal CadA contains more charged amino acids than that of Helicobacter CadA (Fig. 4). The differences in the arrangement of the charged amino acids were found in other regions of these two enzymes.

For example, there are two lysine residues found in TM3 in

staphylococcal CadA, however, no lysine or other positively charged amino acid are found in this domain of Helicobacter CadA. Four

positively and one negatively charged amino acids are found in C-terminal end of staphylococcal CadA while only two positively charged amino acids are found in that of Helicobacter CadA.

An unexpected result was found in this topological study, that is, the neither chimeras G388P nor G388L displayed fusion enzyme activity (Table IV). The possible reason for that may be due to the circumstance that the reporter enzymes were fused immediately downstream of the TM VI. It was known that phoA and lacZ fusions tend to introduce biases into membrane protein topology analysis, especially when positively charged residues in the amino acid sequence downstream of the TM are absent in the fusion protein (Frank et al., 1999). Similarly, the reporter enzymes in these two clones were possibly embedded into the membrane because the lysine anchor at the amino acid position 392 was absent in these chimeras.

Putting all these information together, a detail CadA topography is thus to establish (Fig. 8).

Furthermore, we have demonstrated the presence of these chimera proteins in our study. Since there might be a possibility that no enzyme activity measured is due to no production in these chimera, therefore, it will be very important to clarify if these chimera proteins are produced in the study. As shown in Fig. 9, proteins prepared from those chimera clones were subject to SDS-PAGE analysis and the bands of the target proteins were visualized by using immunoblotting analysis.

Demonstrating that all these chimera proteins were produced in each particular clones, however, some with a small amount of protein

production, and some are larger (Fig. 9). It was noted that the periplasmic fusions with these reporter enzymes expressed less hybrid proteins than those of cytoplasmic fusions. This observation has also been previously

53

reported when the fusions with higher alkaline phosphatase activities are sometime toxic to cells, especially the fusions at the third periplasmic location (Allard and Bertrand, 1992). Similarly, the chimera protein with β-galactosidase is also toxic when secreted to the periplasm of E. coli (Snyder and Silhavy, 1995). Some other studies shown that

transmembrane-directed proteolysis of a membrane-spanning fusion protein causes periplasmic cleavage by a bacteriophage assembly protein (Guy-Caffey and Webster, 1993), a newly finding of a proteolytic

transmembrane signaling pathway, the β-lactamase regulatory system, might be also involved in the periplasmic cleavage (Zhang et al., 2001).

Therefore, it would not be surprised that these periplasmic fusions generate proteins slowly, but degrade them rapidly.

In some of these chimera proteins, bands of lower molecular weights were observed and these might be due to the degradation of these fusion proteins. As it has been suggested that an E. coli FtsH could

initiates proteolysis reaction to rapidly degrade fusion proteins if they were not tightly folded (Kihara et al., 1999). Therefore, along with those predicted fusion protein bands can be seen by Western blotting analysis, some degraded protein bands in each chimera sample preparation were also observed (Fig. 9). These data strongly demonstrated that all the chimera protein were made and their unable to detect either alkaline phosphatase or β-galactosidase enzyme activity in certain constructs were not due to the problem of protein production, but the enzyme orientation on the membrane instead. Among the chimera proteins analyzed in this study, we found that proteins produced by L130P, E181P and L109L clones displayed a higher molecular weight than they should be expected (Fig. 9). This latter observation could be explained as that the chimera

proteins maintaining native conformations and binding less SDS in the electrophoresis buffer as previously described (Guan et al., 1999). The mobility retardation might also result from the formation of disulfide bonds between the alkaline phosphatase (Derman and Beckwith, 1991).

As mentioned previously, more detail characterizations would be necessary to provide a close look into the common signatures in

CadA-related CPx-type ATPases, which including the metal binding property,phosphatase activity and aspartyl kinase activity of the enzyme.

CadA-related CPx-type ATPases, which including the metal binding property,phosphatase activity and aspartyl kinase activity of the enzyme.

在文檔中 對金黃色葡萄球菌上抗鎘腺核甘三磷酸酶穿膜 (頁 58-0)