Indian Journal of Biotechnology Vol I, October 2002, pp 386-392

Diagnostic Tests on the Mode of Ligand Binding to ~roteins: Application to Zymomonas mobilis Strains

EM PapamichaeI'*, A-I Koukkou l, E Douka l

, G Vartholomatos l, K A Masavetas2 and C Drainas '

I Department of Chemistry, University of Ioannina, Ioannina 451 10, Greece

2 Department of Chemical Engineering, Section of Materials Science and Engineering, Laboratory of Physical Chemi stry, Iroon Polytechniou, National Technical University, 9 Zografou, Athens 157 80, Greece

Received 20 May 2002; accepted 20 July 2002

The occurrence of biphasic responses when kinetic data, which describe the binding of ligands to proteins are outlined in Eadie - Hofstee plots are not uncommon as in several cases of glucose uptake of Zymolllonas lIIobilis strain A TCC 10988 and its derivative CU1Rifl/c\one. In this work the authors report two novel diagnostic tests, which can be used easily and routinely in distinguishing the two possible cases. In the first case, two proteinaceous components are involved having one binding site per protein molecule and in the second case a single proteinaceous component having two binding sites per protein molecule is involved. These tests are based on the evaluation of either thefractal dimension and/or the coefficients of a virial expansion, of the Michaelis-Menten equation.

Keywords: Zymomonas mobilis, diagnostic, fractal, virial,glucose uptake

Introduction

The ability of Zymomonas mobiLis to grow on elevated glucose concentrations using the derivative CV I Rif2 of strain A TCC 10988 was investigated (Douka et aL, 1999). Although, CVIRif2 showed more than 20 hrs lag period when transferred from a 0.11 M to 0.55 M glucose liquid medium, however, it grew normally on elevated concentrations of other sugars. CVIRif2 cells grew without any delay on 0.55 M glucose on which wild type cells had been incubated for 3 hrs and removed at the beginning of their exponential phase. At that time, the authors suggested that in Z. mobilis a diffusible heat-labile proteinaceous factor, transitionally not present in 0.55 M glucose of CVIRif2 cultures, triggers growth on 0.55 M glucose. Biochemical analysis of glucose uptake and glycolytic enzymes implied that glucose assimilation was not directly involved in that phenomenon. However, in several cases glucose update data showed biphasic responses In

Eadie- Hofstee plots.

* Author for correspondence: Tel: +30-651098395; Fax: +30-651-047832 E-mail: epapamic@cc.uoi .gr

It has been accepted that Z. mobilis transports glucose by means of a low-affinity, high-velocity glucose-facilitated diffusion system based on GLF protein (Parker et aL, 1995; DiMarco & Romano, 1987; Parker et aL, 1997). On the other hand, Walsh and coworkers (Walsh et aL, 1994) based on the biphasic nature of Eadie -Hofstee plots obtained from uptake experiments using Saccharomyces cerevisiae, postulated that glucose transport could be composed of one or more similar transporter proteins. They suggested that poor data analysis may contribute to the lack of progress in the area of the elucidation of glucose transport.

In this paper the authors present new results helpful in the diagnosis of number of possible transporter proteins of the glucose transport in Z. nzobiLis. Consequently, they applied a conventional methodology, by using the strain ATCC 10988 as well as its CV 1 Rif2 mutant, and collected series of glucose uptake experimental data, which were analyzed kinetically by applying two novel diagnostic tests. Materials and Methods

Strains and growth conditions, estimation of glucose concentrations, bacterial conjugation, and

PAPAMICHAEL et af: DIAGNOSTIC TESTS ON LIGAND BINDING TO PROTEINS 387

DNA methods and sequencing have been described in detail (Douka et al, 1999). The glucose uptake assays protocol was as follows: cells were harvested at the mid-exponential phase, washed with phosphate buffer (lOO mM, pH 6.5) and re-suspended in the same buffer essentially as described by others (Walsh et ai, 1994). Glucose uptake was measured using D-[U-14C]glucose (Amersham, England, 291 mCi/mmol) at concentrations ranging from 0.25 to 50 mM. Z. lIlobilis cells (50 Ill) and 5-fold-concentrated, radiolabelled glucose (12.5 Ill) were preincubated separately at 20°C, mixed together to give the appropriate glucose concentration and vortexed immediately. The uptake was stopped by adding 10 ml phosphate cold buffer (lOO mM, pH 7.5, -2.5°C) containing 500 mM unlabelled glucose. Cells were immediately filtered and washed with 10 ml of the same buffer. The uptake rate was expressed as nanomoles of glucose taken up per min per mg of total protein.

Results and Discussion

Table 1 depicts the whole series of glucose uptake data (Douka et al, 1999). The one kinetic component Eg. (I), analogous to the Michaelis-Menten one, failed to fit all collected series of glucose uptake experimental data by nonlinear regression. In all cases, in the second column of Table 1, glucose uptake data exhibited biphasic responses in Eadie­Hofstee plots, and thus additional parameters were determined by nonlinear curve fitting of Eqs. (2) and (3) to these data.

v - Vmax [5] (uptake) - Kill + [5]

v [5] max

2

K 111 [5]

I + --~ + -------[5j K

1112

Vmax1 [5] Vmax2 [5] v - +---=---

(uptake)- Kill + [5] Kill + [5] I 2

(I)

(2)

(3)

Eg. (2) is valid when one molecule of proteinaceous transporter comprises two binding sites

Table I-A summary of glucose uptake data exhibiting Michaelis-Menten like behaviour. and other biphasic kinetics in

Eadie-Hofstee plots

Michaelis-Menten behaviour Biphasic Eadie-Hofstee plots

ATCC 10988 at 2% Glucose ATCC 10988 at S% Glucose CU IRif2 at 2% Glucose 10988/pLAFRS at 5% Glucose CUIRif2 at 5% Glucose CUIRif2/c1one at 2%

Glucose 10988/pLAFRS at 2% :U I Rif2/c1one at S% Glucose Glucose

A genomic library from ZYlllolllol1a.l' 1II0bi/is CP4 was transferred to strain CUIRif2 by bacterial conjugation. Two clones resistant to O.SS M glucose were selected from 600 transconjugant colonies. One of them was further subcloned in pLAFRS. The resulting pLAFRS recombinants were transformed and transferred by conjugation to CU I Rif2. as a 9.S-kb fragment complementing the glucose sensitive phenotype of CU I Rif2. The complete restriction analysis. subcloning of the 9.S-kb fragment and complementation data in the glucose sensitive mutant are described and di scussed by Douka et al (1999).

one of low and another of high affinity, available for two glucose molecules, respectively. Correspon­dingly, Eq. (3) is valid when each molecule of two different proteinaceous transporters, comprises one binding site available only for one molecule of glucose. In Eqs (2) and (3), Kill and Kill are the

I 2

apparent affinity constants, and VlIlax and Villa\" are I " 2

the respective maximal rates of glucose uptake; (S) is the concentration of glucose. In all fitting procedures, the Least Squares criterion of convergence was used; in addition, robust weighting and four statistical criteria were applied (Theodorou et al, 2000) in order to decide which equation best fitted each series of experimental data. All uptake experimental data. appearing in the second column of Table I, fitted equally in both the Egs (2) and (3). Therefore, based on the results from these fitting procedures, one cannot distinguish whether one transporter having two binding sites (one of low and another of high affinity) exists or two glucose transporters each of them having one binding site exist.

To resolve this uncertainty, the authors developed two suitable diagnostic tests, which were applied to the above experimental data. As it has been pro­posed earlier (L6pez-Quintela & Casado, 1989; Lymperopoulos et al, 1998) there are alternative ways to study kinetically the binding of Iigands to proteins, although, both of the previous publications deal with

388 INDIAN J BIOTECHNOL, OCTOBER 2002

enzymic reactions. However, their results can be easily applied to any protein-ligand interactions. In the first publication (L6pez-Quintela & Casado. 1989), authors proposed a fractal approach to the Michaelis-Menten equation of the form appeared in Eq. (4). Alternatively, another form that of Eq. (5) has been proposed (Lymperopoulos et ai, 1998), which is based on the virial expansion of the Michaelis-Menten equation, and it can replace all previous equations dealing with enzyme kinetics.

V(any) = Veff [Sl2 - D)

max

K eff + [S1 m

0 Cl) en -C)

E -:::l (!) en ~ 0 E c:::

0 .3

0.2

0.1

0

0.3

0 .2

0.1

0

(4)

A

0 5 10 15 20

B

0 10 20 :;0 40

[S]mM

(5)

In Eq. (4), 0 is the fractal dimension, and Vel! and II/ax

Kejf are the effective individual parameters m

describing the global character of the studied mechanism; in Eq. (5) A. A. are the virial

2 3

coefficients (Lymperopoulos et ai, 1998; Laidler & Meiser, 1982).

Based on the above two formu lae (4) and (5) the authors investigated the values of the fractal

o

25 30 35 40 45 ~

[5] mM

C

r 50 0 10 20 :;0 40 50

[S]mM

Fig. I-ATCC 10988 at 5% Glucose: a, Best fit of uptake data by Eq. (2), where VIIU/x = 0.05 ± 0.03, VII/OX = 0.28 ± 0.02, Kill = 0.70 ± I 2 I

0.84, and Kill = 16.82 ± 7.43 (at 95 % confidence); .b, Best fit of the same uptake data by Eq. (4), where D = 0.66 ± 0.05 (at 95% 2

confidence); c, Best fit of the same uptake data by Eq. (5), having two virial coefficients, a first negative virial coefficient whose absolute value was estimated as 30 ± 1.8 times larger (at 95% confidence) than that of the second positive one (at 95% confidence).

PAPAMICHAEL et at: DIAGNOSTIC TESTS ON LIGAND BINDING TO PROTEINS 389

dimension D, and the number and the sign of the virial coefficients as well as the relevance of their magnitudes. Moreover, in order to confirm the proposed tests they performed appropriate computer simulations, by using appropriate software (Evmiridis & Papamichael, 1991), and also applying their experience on this matter. Finally, the authors found out a statistical difference at 95% level between Eqs, (2) and (3): (a) in the estimated fractal dimension D, and (b) in the relevance of the magnitude of the first virial coefficient relative to the second one.

The fractal dimension, which corresponds to

0.4 A

0 Q) (/) -C)

0 E -::::J 0 5 10 15 20 25

Eq. (2) was estimated as D = 0.75 ± 0.16 and that of Eq. (3) as D = 0.50 ± 0.09. Accordingly, it was found that an equation similar to (5), having two virial coefficients, the first of which was estimated as negative and the second as positive, could replace Eqs, (2) and (3). The absolute value of first virial coefficient was estimated as 30 ± 0.9 times larger than that of the second one, in case of Eq. (2), and as 22952 ± 100 times larger than that of the second one, in case of Eq. (3).

By fitting Eq. (4) to whole series of experimental data of the second column of Table 1, a mean value of

30 35 40 45 50 C!) (/) [S] mM ~ B C ~ 0.4 ..e c:

0.3

0.2

O. t

0

0 10 20 30 40 50 0 10 20 30 40 50 [S] mM [S] mM

Fig. 2-CU I Rif2/c\one at 2% Glucose: A.Best fit of uptake data by Eq. (2), where Vlllax = 0.14 ± 0.06, Vlllax = 0.47 ± 0.12, Kill = 1.82 ± I 2 I

0.14. and Kill = 21.68 ± 1.22 (at 95 % confidence); B, Best fit of the same uptake data by Eq. (4), where D = 0.79 ± 0.12 (at 95% 2

confidence); C, Best fit of the same uptake data by Eq. (5), having two virial coefficients, a first negative virial coefficient whose the absolute value was estimated as 78 ± 11 times larger (at 95% confidence) than that of the second positive one (at 95% confidence).

390 INDIAN J BIOTECHNOL, OCTOBER 2002

0.4 A

0 .3

0.2

..

0.1

0 CD en - 0 C)

E -~ 0 5 10 15 20 25 30 35 40 45 50 C) [S] mM en ~ 'f' B C 0 ci E c:

/4')

0

N 0 • 0

0

0 10 20 30 40 50 0 10 20 30 40 50 [5] mM [5] mM

Fig. 3--CU I Rif2/c1one at 5% Glucose: A, Best fit of uptake data by Eq. (2), where Vlllax = 0.07 ± 0.32, Vlllax = 0.38 ± 0.23. Kill = 2.05 ± I 2 I

9.88, and Kill = 19.90 ± 35.71 (at 95 % confidence) ;. B, Best fit of the same uptake data by Eg. (4), where D = 0.78 ± 0. 16 (at 95% 1

confidence); C, Best fi t of the same uptake data by Eg. (5), having two virial coefficients, a first negative viria l coefficient whose absolute value was estimated as 8 1 ± 13 times larger (at 95% confidence) than that of the second positive one (at 95% confidence).

o = 0.73 ± 0.22 (at 95% confidence) for the fractal dimension was estimated. Additionally, by fitt ing Eq . (5) with two virial coefficients, to whole series of experimental data of the second column of Table 1, a negative first virial coefficient was estimated whose absolute value was 76 ± 12 (at 95% confidence) times larger that that of the second positive one. Hence, it is more likely to conclude that for the system under investigation there is only one proteinaceous factor, which functions as a glucose transporter having two

binding sites; the binding of one glucose molecule onto the first site enhances the binding of a second glucose molecule. This is the first report on the diagnosis between two kinetic equations having the same number of parameters and giving the same biphasic plot when one tries to rearrange their nonlinear forms.

Figures 1-3 depict selected examples from the fitting of cases presented in Table I, where Eq. (2) and Eq. (5) having two virial coefficients, fitted the

PAPAMICHAEL et al: DIAGNOSTIC TESTS ON LIGAND BINDING TO PROTEINS 391

0.7!: A

0.50

0.25

(.) 0 .,

~ C)

0 0.5 E 1.5 2 2.5 3 -::l [S] mMX10 (!) B rn 0.75 c .!! 0 E 0.50 C

•

0.25

0

0 1 2 o 1 2

[S] mMX10 [S] mMX10

Fig. 4-An example from simulated uptake data; the case of two different proteinaceous transporters, comprising only one binding site, for one Glucose molecule: A, Best fit of uptake data by Eq. (3), where V

lllax = 0.14 ± 0.01 , V

lllax = 2.75 ± 1.90, Kill = 3.59 ± 1.88. and

I 2 J

Kill = 1727.40 ± 14.41 (at 95 % confidence); B, Best fit of the same uptake data by equation (4), where D = 0.55 ± 0.16 (at 95% 2

confidence); C, Best fit of the same uptake data by Eq. (5), having two virial coefficients, a first negative virial coefficient whose the absolute value was estimated as 22285 ± 113 times larger (at 95% confidence) than that of the second positive one (at 95% confidence).

corresponding experimental data for glucose uptake. Fig. 4 depicts an example where simulated data were best fitted in Eq. (3) and Eq. (5) having two virial coefficients. In all figures, the estimated values of Vmax are given in nmoles of glucose/mg/sec, the values of Km in mM, and the parameter D and the virial coefficients are dimensionless numbers.

References

DiMarco A A & Romano A, 1987. D-Glucose transport system of Zymo/llonas /IIobilis mutant with delayed growth on high glucose concentrations. Appl Environ Microbiol, 49, 151-157.

Douka E et ai, 1999. A Zymomollas mobilis mutant with delayed

growth on high glucose concentrations. J Bacteriol. 181,

4598-4604.

Evmiridis N P & Papamichael E M, 1991. Investigati on of the

error structure of the calibration curve for periodate

determination by flow-injection analysis and

chemiluminescence detection. Chem illfell Lab Systems. 12,

39-47.

Laidler K J & Meiser J H, 1982. Physical Chemistry, The

Benjamin/Cummings Publishing Co. Inc. Pp 2-41.

L6pez-Quintela A-M & Cas ado J, 1989. Revision of the

methodology in enzyme kinetics: A fractal approach. J Theor

Bioi, 139, 129-139.

392 INDIAN J BIOTECHNOL, OCTOBER 2002

Lymperopoulos K et ai, 1998. A formulation of different

equations applied in enzyme kinetics. J Sci Ind Res, 57, 604-

606.

Parker C et al. 1995. Characterization of the Zymomonas mobilis

glucose facilitator gene product (glf) in recombinant

Escherichia coli: Examination of transport mechanism,

kinetics and the role of glucokinase in glucose transport. Mol

Microbiol. 15, 795-802.

Parker C et aI, 1997. Kinetics phosphorylation influence on cometabolism by Zymomonas Microbiol. 63, 3519-3525.

of sugar transport and glucose and fructose

mobilis. Appl Ellviroll

Theodorou L G et ai, 2000. Insight into the catalysis of hydrolysis of four newly synthesized substrates by papain: A proton inventory study. Biochemistry, 13, 3996-4004.

Walsh M C et ai, 1994. Affinity of glucose transport in Saccharomyces cerevisiae is modulated during growth on glucose. J Bacteriol. 176,953-958.