ANALYTICAL

Analytical Biochemistry 330 (2004) 219–226

BIOCHEMISTRY

www.elsevier.com/locate/yabio

A pH-sensitive assay for galactosyltransferase

Chenghua Deng and Rachel R. Chen*

Chemical Engineering Department, Virginia Commonwealth University, 601 W. Main Street, Richmond, VA 23284, USA

Received 9 December 2003

Available online 20 May 2004

Abstract

We report here a new pH-indicator-based assay for galactosyltransferase. The method is simple and fast, requires no specialized

equipment, labeled substrate, or other expensive materials, and is thus expected to have broad applications including automated

high-throughput screening. The method is based upon the detection of absorbance change of a pH indicator, phenol red, in response

to proton release that accompanies the galactosyltransferase-catalyzed galactose transfer. The assay was used to compare three

galactosyltransferases in our collection. As demonstrated here, subtle differences in substrate specificity were readily discerned with

this sensitive method. All three enzymes accept both N -acetylglucosamine and glucose as acceptor but the relative activity varies

with the origin of the enzyme. The method was demonstrated to be useful in the initial characterization of recombinant galacto-

syltransferase from crude cell extract. Optimal metal cofactor Mn2þ concentration and temperature were determined with the

method. Overall, the method offers a great improvement over current methods in reducing time and material consumption. It is the

first pH-sensitive method for galactosyltransferase. The principles of using pH indicator in galactosyltransferase assay should be

applicable to other glycosyltransferase enzymes.

� 2004 Elsevier Inc. All rights reserved.

Keywords: Galactosyltransferase; Proton release; Phenol red; pH-sensitive assay; pH indicator

Galactosyltransferases are a family of enzymes

catalyzing galactose transfer from a donor substrateUDP-galactose to an acceptor molecule. Specifically,

b-1,4-galactosyltransferase (GalT1)1 catalyzes the trans-

fer of galactose from UDP-galactose to N -acetylglucos-

amine (Scheme 1). The disaccharide product of the

reaction, N -acetyllactosamine, is the core structure of

glycans associated with many important biological

processes and human diseases. For example, it occurs in

sialyl Lewisx, an oligosaccharide associated with in-flammation processes, and in Gal(1,3)bGal(1,4)bGlc-

NAc associated with hyperacute rejection of xenografts

[1]. The biomedical relevance of the enzyme has, in

* Corresponding author. Fax: 1-804-828-3846.

E-mail address: rchen@vcu.edu (R.R. Chen).1 Abbreviations used: UDP, uridine diphosphate; GlcNAc, N -acetyl

glucosamine; Glc, glucose; Gal, galactose; NAD, nicotinamide adenine

dinucleotide; HPLC, high-performance liquid chromatography; GalT,

b-1,4-galactosyltransferase (from bovine milk); IPTG, isopropyl

thiogalactopyranoside; OD, optical density.

0003-2697/$ - see front matter � 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.ab.2004.03.014

recent years, stimulated research on this enzyme and its

associated glycans [2–5]. A reliable, sensitive assay ofthis enzyme is essential to all these activities.

Palcic and Sujino [6] thoroughly reviewed current

assay methods for glycosyltransferases. Radiochemical

assays are the most frequently used due to their high

sensitivity. Immunological methods [7] are also sensitive

and are able to identify reaction products, but they re-

quire very specific antibodies or lectins, which are ex-

pensive or, worse, not readily available in most cases. Aspectrophotometric assay developed in 1970s requires

two additional enzymes and is based on detection of

reaction product UDP (Scheme 1) by coupling the gly-

cosyltransferase-catalyzed reaction with two other en-

zymatic reactions that involves NADH/NAD cofactors

[8,9]. Various chromatographic methods developed for

glycosyltransferase assay can combine product identifi-

cation with enzyme assay, but they require substratefluorescence labeling or special detection methods [10].

Mass-spectrometry-based assays [11] are powerful with

regard to speed and accuracy but are not widely used

Scheme 1. Reaction catalyzed by GalT1.

220 C. Deng, R.R. Chen / Analytical Biochemistry 330 (2004) 219–226

due to the high cost of the instrument. Thesemethods and

other more recent modifications [12], requiring substrate

labeling, high instrument investment, or additional en-

zymes or antibodies, are all based upon the detection ofthe substrate consumption of the formation of oligosac-

charide product or nucleotide product. Interestingly,

proton release that accompanies the glycosyltransferase-

catalyzed reactions has never been exploited in the

enzyme assay. This is surprising as pH change due to

proton release of an enzymatic reaction is a feature that

can be conveniently used in assay development, and in-

deed it is used for assays for a wide range of enzymesincluding kinase [13], lipase [14], and phospholipase [15].

In our research focusing on biocatalytic synthesis of

complex carbohydrates, we developed an HPLC method

based on ion-exchange chromatography and electro-

chemical detection [16]. It afforded reliable assay but

was very time consuming. In developing various enzy-

matic synthesis strategies including comparing various

glycosyltransferases and investigating numerous bio-catalytic conditions, we realized that we need a much

faster assay. Therefore, we undertook this study to de-

velop a pH-sensitive assay for b-1,4-galactosyltransfer-ase and to illustrate the methodology that is expected to

be applicable to other galactosyltransferases and other

glycosyltransferases.

Materials and methods

Materials and equipment

All the reagents including sodium phosphate (mono-

basic and di-basic), phenol red, N -acetylglucosamine,

UDP-galactose, UDP-glucose, galactose, glucosamine,

lactose, glucose, N -acetylgalactosamine, isopropyl thi-ogalactopyranoside (IPTG), and a-lactalbumin are from

Sigma. Galactosyltransferase from bovine milk is also

from Sigma. The absorbance at 557 nm of phenol red

was recorded with a Perkin–Elmer Lambda 40 UV–Vis

spectrometer.

HPLC method

Lactose synthesis catalyzed by GalT was detected by

a high-pH anion exchange chromatography system with

pulsed amperometric detection (Dionex). A carbohy-

drate column PA-10 (250� 4mm) was used in the

analysis and 40mM NaOH was used to elute the

product followed by a regeneration of the anion ex-

change column with 200mM NaOH at the end of eachrun. A typical run takes about 45min.

Enzyme preparation

Two recombinant enzymes were expressed in Esche-

richia coli, the genes were originated from Neisserria

meningitidis andHelicobacter pylori and designated LgtB

[17] and HP0826 [18], respectively. Both enzyme con-structs were gifts from Dr. Wakarchuk of the National

Research Council of Canada. The crude enzymes (LgtB

and HP0826) were prepared according to the procedure

of Blixt et al. [3] with some modifications. A glycerol

stock of E. coli strains carrying appropriate plasmid was

inoculated into LB (10 ll stock/ml LB) with 150 lg/ml

ampicillin and incubated at 37 �C overnight. Then 1ml of

the overnight culture was transferred into 100ml LB with150 lg/ml ampicillin and incubated at 37 �C until the

OD600 reached between 0.6 and 0.8 (2–2.5 h). Protein

expression was induced by adding IPTG to a final con-

centration of 0.4mM and the culture was further incu-

bated for 6 h. Cells were harvested by centrifugation. The

pellet was resuspended in 1ml phosphate buffer (2mM,

pH 8.0). Enzymes were released by sonication. The pellet

was removed by centrifugation, and supernatant (con-taining crude enzyme) was collected and stored at )20 �Cor used as indicated in this study.

Calibration curve: relationship between proton production

and absorbance of the pH indicator

In a 2mM sodium phosphate buffer (pH 8, 1ml)

containing 0.01mM phenol red, 0.1mM MnCl2, 10mMN -acetylglucosamine, and LgtB or HP0826 (50 ll), dif-ferent amounts of hydrochloric acid (10mM) were ad-

ded to final concentrations of 0.1, 0.2, 0.3, 0.4, 0.5, 0.6,

0.7, and 0.8mM, and the OD557 was recorded. A

quantitative linear relationship between proton con-

centration and absorbance was established.

Enzyme assay

For this assay, 0.01mM phenol red, 0.1mM MnCl2,

10mM N -acetylglucosamine, and an enzyme of appro-

C. Deng, R.R. Chen / Analytical Biochemistry 330 (2004) 219–226 221

priate volume (typically 50 ll of LgtB, or 50 ll ofHP0826, or 20 ll of GalT) were mixed with phosphate

buffer (2mM, pH 8). The reaction was started by adding

UDP-galactose to a final concentration of 2mM, and

the absorbance at 557 nm was recorded for each sample

at 10-s intervals for a total of 5min. The total volume is

1ml. All the measurements were carried out at 30 �Cexcept as indicated otherwise. The activities of enzymes

were calculated from the calibration curve.To compare substrate specificity of the three enzymes,

additional acceptor substrates including glucose, galac-

tose, lactose, N -acetylgalactosamine, and glucosamine

were used as acceptor. For donor substrate specificity,

UDP-glucose and UDP-galactose were used, and

N -acetylglucosamine was used as acceptor. All mea-

surements were done in triplicate and values averaged.

For reactions catalyzed by GalT, 0.2mg/ml a-lactalbu-min was added except as indicated otherwise.

Effect of temperature on the activity of enzyme

Optimal temperatures were determined under condi-

tions as described above, except that measurements were

taken at different temperatures (20, 30, or 37 �C).

Effect of a-lactalbumin on enzyme activity

Influence of a-lactalbumin on enzyme kinetics was

determined under essentially the same conditions as

described above, except that a-lactalbumin was added to

the mixture to a final concentration of 0.2mg/ml.

Effect ofmanganese on the activity of galactosyltransferase

Optimal cofactor concentration was determined un-

der essentially the same conditions as described above,

except that the final concentration of MnCl2 was

changed to 0, 0.05, 0.2, or 0.5mM.

Fig. 1. Calibration curve: relationship between proton concentration

and absorbance at 557 nm. The mixture contains 2mM phosphate, pH

8.0, 0.01mM phenol red, 0.1mMMnCl2, 10mM N -acetylglucosamine,

and 50ll LgtB. OD557 was measured after addition of different

amounts of hydrochloric acid (0–0.8mM).

Results and discussion

Method development

When choosing an appropriate pH indicator to use in

the assay development, we considered the pH optimal of

galactosyltransferase, which is close to 8.0, and the pH

range for the color change of an indicator. We chose

phenol red since its pKa is 7.4 and it transits from red toyellow with associated absorbance change within a pH

range of 8.4–6.8, coinciding with a pH range within

which galactosyltransferase is expected to be most ac-

tive, though not necessarily at its optimal. Phenol red

has a peak absorbance at 557 nm; this wavelength is

used in all of our subsequent experiments. Another

important consideration in assay development is the

choice of pH buffer concentration. A higher bufferconcentration is not desirable as it renders the method

insensitive to pH change, the basis of the detection.

Buffer concentration too low is problematic as absorp-

tion of CO2 from air will cause fluctuations in pH or

proton measurement. Generally, pH buffer concentra-

tion in a pH-indicator-based assay should be kept as low

as possible. After an initial investigation, a buffer con-

centration of 2mM was chosen. Earlier studies [13]show that a pH buffer with a pKa as close to the pKa of

the indicator as possible gives the best result. For this

reason, we chose phosphate buffer (pKa 7.2) in our assay

development, which was found to be satisfactory.

A galactosyltransferase-catalyzed reaction is a pro-

ton-producing reaction; thus pH will decrease as reac-

tion proceeds. If the pH falls within the color-change

range, the color of phenol red changes from red to yel-low, with a corresponding absorbance decrease. The

absorbance decrease, correlated with the proton con-

centration increase by an appropriate calibration curve,

can then be used as the basis for quantitative analysis of

galactosyltransferase.

Titration using HCl was used to establish the rela-

tionship between proton concentration and absorbance

of the indicator (details are described under Materialsand methods). As seen in Fig. 1, absorbance measured

at 557 nm is linearly correlated with proton concentra-

tion from 0 to 0.8mM, which covers the range of proton

concentration of all experiments in this study. Note that

the calibration curve was generated with essentially the

same enzyme mixture as that used in the actual assay

omitting only the donor substrate.

A typical time trajectory of absorbance change cor-responding to LgtB-catalyzed proton release is shown in

Fig. 2. Using the calibration curve shown in Fig. 1, the

proton concentration change corresponding to the

absorbance change was calculated and plotted as a

Fig. 2. Change of absorbance at 557 nm with time for LgtB. The en-

zyme was assayed in 2mM phosphate, pH 8.0, 0.01mM phenol red,

0.1mM MnCl2, 10mM N -acetylglucosamine, and 50 ll LgtB. The

reaction was started by adding 2mM UDP-galactose, and measure-

ments were taken at 557 nm for 5min with a time interval of 10 s. The

measurement was repeated in triplicate and the values were averaged.

The curves represent the average and standard deviation. No UDP-

galactose was added in Blank.

Fig. 3. Release of proton as a function of time from an LgtB-catalyzed

reaction. The enzyme was assayed in 2mM phosphate, pH 8.0,

0.01mM phenol red, 0.1mM MnCl2, 10mM N -acetylglucosamine,

and 50 ll LgtB. The reaction was started by adding 2mM UDP-ga-

lactose, and measurements were taken at 557 nm for 2min with a time

interval of 10 s. The measurement was repeated in triplicate and the

values were averaged.

Fig. 4. Change in absorbance as a function of time with different en-

zyme input. The changes in OD557 were recorded in 2mM phosphate,

pH 8.0, 0.01mM phenol red, 0.1mM MnCl2, 10mM N -acetylglucos-

amine, and LgtB (50, 10, 5, and 0 ll). The reactions were started by

adding 2mM UDP-galactose and measurements were taken at 557 nm

for 5min. Data were collected at each 10-s time interval. The mea-

surement was repeated in triplicate and the values were averaged.

222 C. Deng, R.R. Chen / Analytical Biochemistry 330 (2004) 219–226

function of time (Fig. 3). A linear regression was per-

formed (R2 ¼ 0:9987) and the slope was estimated to be

0.0953. The enzyme activity was defined as the amount

of enzyme required to produce 1 lmol proton per min-

ute. The particular example (enzyme LgtB) shown in

Figs. 2 and 3 was found to have activity of about 1.9U

using the following formula:

0:0953mmol=L=min� 1� 10�3 L

0:05ml¼ 1:906lmol=min=ml

¼ 1:906U=ml:

The detection limit of the method was estimated byvarying enzyme volumes used in the assay. Fig. 4 shows

the rate of the reaction decreases with the decrease of

enzyme input. The linear relationship is maintained until

enzyme input decreases to below 10 ll, equivalent to

0.019U. Therefore, the detection limit for this assay

with a 1ml total assay volume is 0.019U. If this assay is

adapted for a 96-well format with a 200 ll total assayvolume, the limit will be reduced to 0.0038U/well.

We compared the method with an established HPLC

method in which N -acetyllactosamine, the product of

galactosyltransferase-catalyzed reaction, was measured

[16]. The activity measured from this assay was consis-

tently about 50% higher than that from the HPLC

method. This difference can be attributed to two main

reasons. First, in this assay, proton production accom-

panying the glycosyltransferase-catalyzed reactions wasused whereas, in the HPLC method, the disaccharide

product N -acetyllactosamine was measured. The differ-

ences in chemical entities measured in the assay and

different instruments used (spectrophotometer versus

electrochemical detector) could be possible sources of

the difference. Second, the HPLC method takes mea-

surement at time points which are on orders of at least

minutes (typically 5, 10, 20, and 30min), whereas ourproton-based assay afforded measurements every 10 s;

thus the measured rate by this method reflects more

closely the initial velocity (typically the first 2min), ra-

ther than the averaged first 30min rate in the HPLC

method. We observed the proton change rate or the

reaction rate decrease with time even within the first

5min of the reaction. In Fig. 1, for example, the rate of

absorption change decreased after the first 2min, indi-cating that the reaction deaccelerates shortly after the

initiation of the reaction possibly because of the well-

known inhibition effect of UDP. Thus it is not surprising

C. Deng, R.R. Chen / Analytical Biochemistry 330 (2004) 219–226 223

to see a lower numerical value in the HPLC assay. Wetherefore believe that the proton-based assay is a more

accurate method as it more truly reflects the initial ve-

locity of the enzymatic reaction.

In the example shown above and elsewhere in this

paper, recombinant enzymes LgtB and HP0826 were

Table 1

Comparison of substrate specificities for LgtB, HP0826, and GalT

Donor Acceptor

a The activities of LgtB, HP0826, and GalT were measured at 30 �C in 2

N -acetylglucosamine, and 50 ll LgtB or 50 ll HP0826 or 20ll GalT. The rea

and the absorbance at 557 nm was recorded for each sample at each 10-s time

and 2 and on the equation in the text. The measurements were done in tripl

used as cell extract without purification. Under theconditions that we used, the cell extract contains 1.36

and 1.15mg/ml total proteins for LgtB and HP0826,

respectively, of which about 20% is the recombinant

enzyme as estimated from our SDS–PAGE gel analysis

(data not shown). Therefore, the recombinant enzymes

Activity (U/ml)a

LgtB HP0826 GalT

1.91� 0.08 1.18� 0.09 6.18� 0.14

1.37� 0.06 0.89� 0.02 19.1� 0.12

0 0 0

0 0 0

0 0 0

0 0 0.47� 0.05

0.39� 0.02 0 0

mM phosphate, pH 8.0, 0.01mM phenol red, 0.1mM MnCl2, 10mM

ctions were started by adding 2mM UDP-galactose or UDP-glucose,

interval for a total of 2min. The activity was calculated based on Figs. 1

icate and the data shown are the average� standard deviation.

Fig. 5. Substrate specificity of LgtB. The changes in OD557 for LgtB to

different substrates were recorded in 2mM phosphate, pH 8.0,

0.01mM phenol red, 0.1mM MnCl2, 10mM N -acetylglucosamine (or

glucose or glucosamine), and 50ll LgtB. The reactions were started by

adding 2mM UDP-galactose (or UDP-glucose), and measurements

were taken at 557 nm for 5min. Data were collected at each 10-s time

interval. No UDP-galactose or UDP-glucose was added in Blank.

224 C. Deng, R.R. Chen / Analytical Biochemistry 330 (2004) 219–226

LgtB and HP0826 represented major components of thecell extract and the compositions of the cell extract were

consistent throughout the study as identical conditions

were used in each step of the enzyme preparation pro-

cess. Furthermore, in all the experiments described in

this paper, adequate controls were used to ensure that

the proton change measured corresponds to the enzyme

activity rather than to other activities in the extract. In

particular, the blank contains the cell extract from thesame preparation and other reagents necessary for the

reaction (details under Materials and methods) except

an acceptor sugar or donor substrate. As seen in Fig. 2,

the change of OD of the blank in the first 2min of the

experiments was minimal compared to the sample in

which UDP-Gal was added, indicating that the change

of absorption and therefore proton concentration is

due to the activity of the galactosyltransferase and thatthe interference of other components in the extract is

minimal.

As in many applications such as biocatalysis

recombinant enzymes are used as crude extract, it is

important to have a quick method for an initial char-

acterization of enzyme properties and a quick screening

method for various process conditions. The ability of

this method to detect galactosyltransferase activity in acrude cell extract is therefore an advantage as it elimi-

nates the need for purification.

Substrate specificities of three different galactosyltransfe-

rases

We used the method to compare three different ga-

lactosyltransferases in our collection. Of the three en-zymes, two were recombinant and were expressed in

E. coli. The genes were originated from N. meningitidis

and H. pylori and designated LgtB and HP0826,

respectively. Both enzymes catalyze galactose transfer

from a donor substrate UDP-galactose to an acceptor

molecule N -acetylglucosamine as shown in Scheme 1.

The third enzyme is a commercial enzyme purified from

bovine milk, here designated GalT, which catalyzesgalactose transfer from UDP-galactose to glucose. The

recombinant enzymes were prepared as described under

Materials and methods.

The substrate specificity of these three enzymes was

investigated with a panel of six acceptors and two do-

nors. As shown in Table 1 and Fig. 5, LgtB and HP0826

showed highest activity with donor UDP-galactose and

acceptor N -acetylglucosamine, whereas the commercialenzyme GalT shows highest activity with glucose as

acceptor and UDP-galactose as donor. These results are

consistent with findings from the literature, indicating

that our method is sensitive and useful in characterizing

substrate specificity. Also consistent with earlier studies,

the enzyme preparation containing LgtB was active with

UDP-glucose, with about 25% activity toward UDP-

glucose (Table 1). This is likely due to the epimerization

of donor sugar as indicated by an earlier study with the

LgtB construct [3] (pers. commun.). The other two en-

zymes show no detectable activity toward UDP-glucose.

All three enzymes accept both N -acetylglucosamine

and glucose as acceptor substrate. GalT also has a low-

level activity toward disaccharide lactose, yet it showsno activity with acceptor monosaccharide galactose.

As demonstrated in this example, the method is sen-

sitive enough to differentiate closely related galac-

tosyltransferases by their relative activities and by their

subtle differences in substrate specificities.

Galactosyltransferase modification by a-lactalbumin

Mammalian galactosyltransferase GalT1 is excep-

tional in that its kinetics can be modified by regulatory

protein a-lactalbumin [4,19]. The bovine commercial

enzyme used in this study is also called lactose synthase.

It synthesizes lactose slowly in the absence of a-lactal-bumin. In the presence of a-lactalbumin, however, the

reaction rate increases dramatically due to the reduced

Km by the regulatory protein. The initial rate of lactosesynthesis catalyzed by GalT with and without a-lactal-bumin was monitored by the developed method. Fig. 6

shows the absorbance change during the first 5min of

synthesis. As illustrated in Fig. 6, in the absence of a-lactalbumin, the reaction barely proceeds. In contrast,

rapid synthesis of lactose takes place as soon as a-lact-albumin is added to the reaction mixture. Lactose for-

mation was monitored using an HPLC method. After24 h, the yield of lactose synthesis based upon UDP-

galactose was over 90%. As expected, a-lactalbumin has

no effect on either LgtB or HP0826. When reactions

Fig. 6. Comparison of the activities of GalT and LgtB with and

without a-lactalbumin. The changes in OD557 for LgtB and GalT were

recorded in 2mM phosphate, pH 8.0, 0.01mM phenol red, 0.1mM

MnCl2, 10mM glucose, 0 or 0.2mg/ml a-lactalbumin, and 50 ll LgtB(20ll GalT). The reactions were started by adding 2mM UDP-ga-

lactose, and measurements were taken at 557 nm for 5min. Data were

collected at each 10-s time interval.

C. Deng, R.R. Chen / Analytical Biochemistry 330 (2004) 219–226 225

were carried out with N -acetylglucosamine, the presence

of a-lactalbumin significantly decreases the reaction rate

catalyzed by GalT (data not shown). a-Lactalbumin,

therefore, is a very effective inhibitor of the enzyme,

inhibiting the transfer of galactose to N -acetylglucos-

amine.The applicability of the method to the detection of

small changes in enzyme kinetics caused by the action of

regulatory molecules, as shown in this example, is a

feature important for drug discovery and enzyme-based

diagnostic assays or other bioassays. Thus we expect

Table 2

Activities of three galactosyltransferases at different temperatures

Temperature (�C)

Activity (U/ml) LgtB (UDP-Gal +GlcNAc)

HP0826 (UDP-Gal+GlcNAc)

Commercial GalT (UDP-Gal +Glc)

The activities of LgtB, HP0826, and GalT were measured in 2mM

N -acetylglucosamine for LgtB or HP0826 (or 10mM glucose for GalT), and

by adding 2mMUDP-galactose and measurements were taken at 557 nm for 2

were calculated according to the method described in the text. The m

average� standard deviation.

Table 3

Optimal Mn2þ concentration

Concentration of MnCl2 (mM) 0

Relative activity (U/ml) LgtB (UDP-Gal+GlcNAc) 0.89� 0.0

HP0826 (UDP-Gal +GlcNAc) 0.67� 0.0

GalT (UDP-Gal+Glc) 7.30� 0.1

The activities were measured in 2mM phosphate, pH 8.0, 0.01mM pheno

20 ll GalT). The concentrations of MnCl2 were 0, 0.05, 0.1, 0.2, and 0.5mM

adding 2mM UDP-galactose and measurements were taken at 557 nm for 2m

to the method described in the text. The measurements were done in triplica

that this sensitive method will have broad applicationssuch as screening of enzyme inhibitors or effectors.

Temperature optimal

Rapid characterization of enzymes with regard to

their temperature optimal is essential to processes em-

ploying enzymes. We tested the ability of this method to

determine temperature optimal in an expeditious man-ner. The assay was carried out with temperature as the

only variable. We found that with this method it re-

quired only about 15min total assay time for each en-

zyme for measurements taken at three different

temperatures. We compared the assay time with a pre-

vious method established in our laboratory based on a

Dionex ion-exchange chromatography [16]. The method

is a significant improvement over the HPLC method asillustrated below. The HPLC method requires about

45min for a single data point and a total of about 4–5 h

for estimation of activity for a single temperature; so if

an assay is conducted with three different temperatures,

a total of 12–15 h of HPLC instrumental time is re-

quired. This represents about 20-fold difference in the

time requirement. Table 2 shows that all three enzymes

have an optimal temperature about 30 �C under theconditions described under Materials and methods.

Manganese cofactor concentration

Galactosyltransferase is known to require Mn2þ as

cofactors [20]. But the optimal concentrations for vari-

ous in vitro applications were not clearly established.

As reported in the literature, they can vary over an

20 30 37

1.78� 0.07 1.91� 0.08 1.54� 0.13

0.89� 0.03 1.18� 0.09 1.05� 0.06

0.18� 0.05 19.1� 0.12 12.0� 0.10

phosphate, pH 8.0, 0.01mM phenol red, 0.1mM MnCl2, 10mM

50 ll LgtB or 50 ll HP0826 or 20ll GalT. The reactions were started

min with a time interval of 10 s at different temperatures. The activities

easurements were done in triplicate and the data shown are the

0.05 0.1 0.2 0.5

9 1.80� 0.10 1.91� 0.08 1.88� 0.06 0.81� 0.06

5 1.02� 0.04 1.18� 0.09 1.33� 0.05 1.11� 0.11

7 12.05� 0.11 19.1� 0.12 17.80� 0.08 16.03� 0.29

l red, 10mM N -acetylglucosamine, and 50 ll LgtB (or 50ll HP0826 or

, respectively, in the five experiments. The reactions were started by

in with a time interval of 10 s. The activities were calculated according

te and the data shown are the average� standard deviation.

226 C. Deng, R.R. Chen / Analytical Biochemistry 330 (2004) 219–226

unusually wide range from 10 lM to 10mM, spanningthree orders of magnitude. We set out to determine the

optimal concentration of Mn2þ with the method devel-

oped. Table 3 shows that both LgtB and GalT have an

optimal about 0.1mM. HP0826, however, has an opti-

mal concentration slightly higher at 0.2mM. Earlier

studies using 10mM Mn2þ in synthesis reactions may

not represent the best conditions under which galacto-

syltransferase should be used.

Conclusion

The developed method based on a pH indicator is

sensitive and very easy to use. The method is useful in

the initial characterization of galactosyltransferase en-

zymes, including determination of optimal temperature

and cofactor concentrations in biocatalytic synthesis

applications, investigation of donor and acceptor sub-

strate specificities in comparison with other similargalactosyltransferases, and screening of potential in-

hibitors and effectors of the enzymes. The method re-

quires only a small volume of enzyme sample, 10–50 llor 0.02U of enzymes in a 1-ml assay format. A typical

run takes 2–5min. Crude enzyme extract can be used

without the need for purification. The method requires

no expensive enzymes or labeled substrates. Overall, this

method is a significant improvement over the currentmethods that employ radiolabeled materials, sophisti-

cated instruments, and time-consuming procedures.

This method is thus widely applicable.

The method is amenable for adaptation as a high-

throughput screening method where a microtiter plate

reader is used. Moreover, with slight adjustment of

conditions and with a suitable choice of a pH indicator

and a suitable buffer, which takes into account the pHoptimal of the enzymes, the method could be easily

adapted for other glycosyltransferase enzymes. It should

be a valuable tool in glycobiology.

Acknowledgments

We thank Dr. Warren W. Wakarchuk and Ms.

Melisa Shur of the National Research Council of Can-

ada for providing E. coli constructs overexpressing LgtBand HP0826 and helpful discussions.

References

[1] E. Bettler, E. Samain, V. Chazalet, C. Bosso, A. Heyraud, D.H.

Joziasse, W. Wakarchuk, A. Imberty, R.A. Geremia, The living

factory: in vivo production of N -acetyllactosamine containing

carbohydrates in E. coli, Glycoconjugate J. 16 (1999) 205–212.

[2] R. €Ohrlein, Glycosyltransferase-catalyzed synthesis of non-natu-

ral oligosaccharide, Top. Curr. Chem. 200 (1999) 227–254.

[3] O. Blixt, J. Brown, M.J. Schur, W. Wakarchuk, J.C. Paulson,

Efficient preparation of natural and synthetic galactosides with a

recombinant b-1,4-galactosyltransferase-/UDP-40-Gal epimerase

fusion protein, J. Org. Chem. 66 (2001) 2442–2448.

[4] E.G. Berger, J. Rohrer, Galactosyltransferase—still up and

running, Biochimie 85 (2003) 261–274.

[5] G.M. Watt, P.A.S. Lowden, S.L. Flitsch, Enzyme-catalyzed

formation of glycosidic linkages, Curr. Opin. Struct. Biol. 7

(1997) 652–660.

[6] M.M. Palcic, K. Sujino, Assay for glycosyltransferases, Trends

Glycosci. Glycotechnol. 13 (2001) 361–370.

[7] L.S. Khraltsova, M.A. Sablina, T.D. Melikhova, D.H. Joziasse,

H. Kaltner, H.-J. Gabius, N.V. Bovin, An enzyme-linked lectin

assay for a-1,3-galactosyltransferase, Anal. Biochem. 280 (2000)

250–257.

[8] D.K. Fitzgerald, B. Colvin, R. Mawal, K.E. Ebner, Enzymic assay

for galactosyl transferase activity of lactose synthetase and alpha-

lactalbumin in purified and crude systems, Anal. Biochem. 36

(1970) 43–61.

[9] S. Gosselin, M. Alhussaini, M.B. Streiff, K. Takabayashi, M.M.

Palcic, A continuous spectrophotometric assay for glyco-

syltransferases, Anal. Biochem. 220 (1994) 92–97.

[10] D.M. Snow, J.H. Shaper, N.L. Shaper, G.W. Hart, Determination

of b-1,4-galactosyltransferase enzymatic activity by capillary

electrophoresis and laser-induced fluorescence detection, Anal.

Biochem. 271 (1999) 36–42.

[11] L. Jobron, K. Sujino, G. Hummel, M.M. Palcic, Glycosyltrans-

ferase assays utilizing N -acetyllactosamine acceptor immobilized

on a cellulose membrane, Anal. Biochem. 323 (2003) 1–6.

[12] B. Abdul-Rahman, E. Ailor, D. Jarvis, M. Betenbaugh, Y.C. Lee,

b-1,4-Galactosyltransferase activity in native and engineered

insect cells measured with time-resolved europium fluorescence,

Carbohydr. Res. 337 (2002) 2181–2186.

[13] E. Chapman, C.-H. Wong, A pH sensitive colorometric assay for

the high-throughput screening of enzyme inhibitors and sub-

strates: a case using kinase, Bioorg. Med. Chem. 10 (2002) 551–

555.

[14] F. Moris-Varas, A. Shah, J. Aikens, N.P. Nadkarni, J.D. Rozzell,

D.C. Demirjian, Visualization of enzyme-catalyzed reactions

using pH indicators: rapid screening of hydrolase libraries and

estimation of the enantioselectivity, Bioorg. Med. Chem. 7 (1999)

2183–2188.

[15] Y. Yao, M.-H. Wang, K.-Y. Zhao, C.-C. Wang, Assay for enzyme

activity by following the absorbance change of pH-indicators, J.

Biochem. Biophys. Methods 36 (1998) 119–130.

[16] D. He, R. Chen, Functional expression of heterologous UDP-

galactose-4-epimerase in Agrobacterium sp. via a broad host-range

expression vector, Biotechnol. Lett. 24 (2002) 1599–1603.

[17] W. Wakarchuk, A. Cunningham, D.C. Watson, N.M. Young,

Role of paired basic residues in the expression of active

recombinant galactosyltransferases from the bacterial pathogen

Neisseria meningitides, Protein Eng. 11 (1998) 295–302.

[18] S.M. Logan, J.W. conlan, M.A. Montero, W.W. Wakarchuk, E.

Altman, Functional genomics ofHelicobacter pylori: identification

of a b-1,4-galactosyltransferase and generation of mutants with

altered lipopolysaccharide, Mol. Microbiol. 35 (2000) 1156–

1167.

[19] F.L. Schanbacher, K.E. Ebner, Galactosyltransferase acceptor

specificity of the lactose synthetase A protein, J. Biol. Chem. 245

(1970) 5057–5061.

[20] D.L. Witsell, C.E. Casey, M.C. Neville, Divalent cation activation

of galactosyltransferase in native mammary Golgi vesicles, J. Biol.

Chem. 265 (1990) 15731–15737.