Heterologous Catalysis of the Final Steps of Tetracycline Biosynthesis by Saccharomyces cerevisiae

Biosynthesis by Saccharomyces cerevisiae

Articles

Heterologous Catalysis of the Final Steps of Tetracycline

Ehud Herbst, Arden Lee, Yi Tang, Scott A. Snyder, and Virginia W. Cornish*

Cite This: https://doi.org/10.1021/acschembio.1c00259

Read Online

ACCESS

Metrics & More

Article Recommendations

sı Supporting Information

ABSTRACT: Developing treatments for antibiotic resistant bacterial infections is among the highest priority public health

challenges worldwide. Tetracyclines, one of the most important classes of antibiotics, have fallen prey to antibiotic resistance,

necessitating the generation of new analogs. Many tetracycline analogs have been accessed through both total synthesis and

semisynthesis, but key C-ring tetracycline analogs remain inaccessible. New methods are needed to unlock access to these analogs,

and heterologous biosynthesis in a tractable host such as Saccharomyces cerevisiae is a candidate method. C-ring analog biosynthesis

can mimic nature’s biosynthesis of tetracyclines from anhydrotetracyclines, but challenges exist, including the absence of the unique

cofactor F₄₂₀in common heterologous hosts. Toward this goal, this paper describes the biosynthesis of tetracycline from

anhydrotetracycline in S. cerevisiae heterologously expressing three enzymes from three bacterial hosts: the anhydrotetracycline

hydroxylase OxyS, the dehydrotetracycline reductase CtcM, and the F₄₂₀reductase FNO. This biosynthesis of tetracycline is enabled

by OxyS performing just one hydroxylation step in S. cerevisiae despite its previous characterization as a double hydroxylase. This

single hydroxylation enabled us to purify and structurally characterize a hypothetical intermediate in oxytetracycline biosynthesis that

can explain structural diﬀerences between oxytetracycline and chlortetracycline. We show that F , a synthetically accessible derivative

of cofactor F₄₂₀, can replace F₄₂₀in tetracycline biosynthesis. Critically, the use of S. cerevisiae for the ﬁnal steps of tetracycline

biosynthesis described herein sets the stage to achieve a total biosynthesis of tetracycline as well as novel tetracycline analogs in S.

cerevisiae with the potential to combat antibiotic-resistant bacteria.

INTRODUCTION

as with any member of the antibiotic armamentarium,

tetracycline use has also led to the spread of antibiotic

■

The development of treatments for antibiotic resistant bacterial

infections is an urgent public health challenge. Indeed, the

resistance, with bacteria having acquired a growing number of

resistance mechanisms including eﬄux pumps, ribosomal

U.S. Center for Disease Control reports over 2 million illnesses

and 20 000 deaths annually in the United States as a result of

protection proteins, rRNA mutations, and enzymatic

degradation.

antibiotic resistance, and the World Health Organization

reports over 700 000 deaths annually worldwide from drug-

The main methods for the production of tetracycline

resistant diseases. Novel antibiotic drug candidates have

antibiotics are biosynthesis and semisynthesis using the

13−15

traditionally been accessed through natural product screening,

total synthesis, semisynthesis, and combinatorial chemistry.

However, the number of new antibiotics approved by the FDA

dropped signiﬁcantly from previous decades, from 29 approvals

in the 1980s to 23 in the 1990s, 9 in the 2000s and 15 in the

original microbial hosts, as well as total synthesis.

Key

subclasses of tetracycline antibiotic analogs remain inaccessible

through these methods, and new methods are needed to access

these analogs. A candidate new method is heterologous

biosynthesis and semisynthesis. The use of the original

tetracyclines host strains, Streptomyces aureofaciens and

Streptomyces rimosus, limits the possible range of tetracycline

010s. To reverse this trend, transformative new technolo-

gies for the discovery of new antibiotics are needed.

Beginning in the 1940s, tetracyclines were used widely for

respiratory, gastrointestinal, and genitourinary tract infections.

Indeed, tetracyclines were among the ﬁrst antibiotic agents

with broad-spectrum activity against numerous Gram-positive

Received: April 6, 2021

Accepted: June 28, 2021

and Gram-negative bacteria, with their mode of action

established as inhibiting bacterial protein synthesis by binding

reversibly to the 30S ribosomal subunit and sterically hindering

aminoacyl-tRNA binding to the ribosomal A-site. However,

XXXX American Chemical Society

ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

Articles

scaﬀolds for semisynthetic modiﬁcations to chlortetracycline,

Scheme 1. Conversion of Anhydrotetracycline to

Tetracycline by Saccharomyces cerevisiae

oxytetracycline, tetracycline, and demeclocycline. The use of

alternative original hosts such as the dactylocycline producer

Dactylosporangium sp. SC 14051 is theoretically possible but

practically limited, because of a very slow growth rate and an

inaccessibility of genetic modiﬁcations. The Myers con-

vergent tetracycline synthesis has been incredibly eﬀective in

generating D-ring tetracycline analogs, including the FDA-

approved eravacycline. However, this powerful synthetic

strategy for forming the AB-ring precursor and then coupling it

with the D-ring precursor across the C-ring limits the overall

number of achievable modiﬁcations at the AB-ring junction

6,17

and the C-ring.

A major heterologous host candidate for the production of

tetracycline analogs is Saccharomyces cerevisiae, which has

proven itself by heterologously producing a variety of natural

products of importance to human health. A key example is the

production of artemisinic acid, a precursor to artemisinin.

More recent examples include the production of hydrocodone,

Δ9-tetrahydrocannabinolic acid (2-carboxy-THC), and pen-

9−21

icillin.

One of the promises in natural product biosyn-

thesis using S. cerevisiae is the access to unnatural analogs of

natural products. Speciﬁcally, heterologous production of

tetracyclines in S. cerevisiae has the potential to yield currently

inaccessible analogs of tetracyclines to combat the resistance to

clinically available tetracycline antibiotics. We envisioned

engineering S. cerevisiae to heterologously express tetracycline

biosynthetic enzymes from original hosts that are genetically

intractable and/or slow growing followed by semisynthetic

modiﬁcations. Heterologous biosynthesis of tetracyclines in S.

cerevisiae would also be instructive for the study of tetracycline

biosynthesis.

The last steps in the biosynthesis of the tetracyclines are

hydroxylation and reduction, starting from an anhydrotetracy-

cline, catalyzed by an FAD-dependent anhydrotetracycline

hydroxylase and an F₄₂₀-dependent dehydrotetracycline

reductase (Scheme 1). In the biosynthesis of oxytetracycline

(

5-hydroxytetracycline), these steps are catalyzed by OxyS and

OxyR, respectively, and in the biosynthesis of chlortetracycline

7-chlorotetracycline), these steps are catalyzed by CtcN and

(

CtcM, respectively. Two key diﬀerences between the two

pathways with regard to these steps are the starting

anhydrotetracycline structure and the number of hydroxylation

steps on that anhydrotetracycline. While OxyS catalyzes two

hydroxylation steps on anhydrotetracycline in the oxytetracy-

cline pathway, CtcN catalyzes only one hydroxylation step on

anhydrochlortetracycline (7-chloroanhydrotetracycline) in the

(

a) The conversion of anhydrotetracycline (1) to oxytetracycline (5)

using puriﬁed proteins OxyS, OxyR, and FGD along with NADPH,

F₄₂₀, and G6P prior to this study. (b) The conversion of

anhydrotetracycline (1) to tetracycline (3) using +OxyS +CtcM

2,23

FNO Saccharomyces cerevisiae whole cells or cell lysates along with

chlortetracycline pathway.

It was hypothesized that

NADPH, F , and G6P in this study; 5(5a)-dehydrotetracycline (2b)

structural diﬀerences between OxyS and CtcN lead to a

diﬀerence in the number of hydroxylation steps between the

and tetracycline (3) were isolated in this study following incubation of

a +OxyS and a +OxyS +CtcM +FNO S. cerevisiae cell lysate with 1,

respectively; the conversion of 5(5a)-dehydrotetracycline (2b) to

oxytetracycline (5) was expected when a +OxyS + OxyR + FNO cell

lysate was used in this study but was not observed. Gray squares

emphasize the carbons at which key chemical transformations occur.

two enzymes.

A unique cofactor to the last steps of the tetracyclines’

^bⁱ^o^s^yⁿ^t^h^e^sⁱ^s^t^h^a^tⁱ^sⁿ^o^tⁿ^a^tⁱ^v^e^t^o^S^.^c^e^r^e^vⁱ^sⁱ^a^eⁱ^s^c^o^f^a^c^t^o^r^F⁴20, a

lactyl oligoglutamate phosphodiester derivative of 7,8-dide-

methyl-8-hydroxy-5-deazariboﬂavin (F ). F is much more

FGD = F₄₂₀-dependent glucose-6-phosphate dehydrogenase; F = 7,8-

synthetically accessible than F₄₂₀. As such, F can act as a

didemethyl-8-hydroxy-5-deazariboﬂavin; FNO = F₄₂₀-NADP oxidor-

eductase from A. fulgidus; G6P = glucose-6-phosphate

substitute for F₄₂₀in some F₄₂₀-dependent reactions in vitro,

but it does not appear to have a redox role in living cells.

F₄₂₀-reducing NADPH dehydrogenase enzymes such as F₄₂₀

anhydrotetracycline to tetracycline, by the heterologous

expression of OxyS, CtcM, and FNO. Importantly, we report

NADPH oxidoreductase (FNO) from Archaeoglobus fulgidus

24,25

can reduce F₄₂₀to its reductively active form, F₄₂₀H₂.

that synthetic F , exogenously supplied to the engineered S.

Here, we describe the use of S. cerevisiae for the ﬁnal steps of

tetracycline biosynthesis, speciﬁcally the conversion of

cerevisiae strain, can successfully replace F₄₂₀in this

biosynthetic pathway. Additionally, we report the character-

ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

Articles

ization of a previously hypothetical intermediate in oxy-

tetracycline biosynthesis that can explain structural diﬀer-

ences between oxytetracycline and chlortetracycline. Tetracy-

cline biosynthesis is enabled in our study because OxyS

performs in S. cerevisiae cells and cell lysates just a single

hydroxylation step on anhydrotetracycline and not two, as

previously reported when puriﬁed enzymes were used. Thus,

our study enhances the understanding of tetracycline biosyn-

thesis and paves the road for total heterologous biosynthesis of

tetracyclines in S. cerevisiae.

RESULTS

■

To carry out the ﬁnal steps of a tetracycline biosynthesis in S.

cerevisiae, namely the conversion of anhydrotetracycline to

tetracycline, we heterologously expressed three enzyme types.

The ﬁrst, an anhydrotetracycline hydroxylase, converts

anhydrotetracycline to dehydrotetracycline. Here, we em-

ployed OxyS from the oxytetracycline pathway as the

anhydrotetracycline hydroxylase. The second, a dehydrote-

tracycline reductase, converts dehydrotetracycline into tetracy-

cline. In this case, we tested three dehydrotetracycline

reductase candidates, OxyR, DacO4, and CtcM from the

oxytetracycline, dactylocycline, and chlortetracycline pathways,

respectively. The third, an F₄₂₀reductase, reduces the cofactor

used by the dehydrotetracycline reductase. Here, three F₄₂₀

reductase candidates from Mycobacterium tuberculosis, Archae-

oglobus fulgidus, and Streptomyces griseus were explored. We

used commercially available anhydrotetracycline as the

substrate in this biosynthesis and synthetic F_oas a

dehydrotetracycline reductase-cofactor instead of its much

Figure 1. Mass spectrometry analysis of anhydrotetracycline

hydroxylation in lysate of S. cerevisiae cells expressing OxyS. Cell

lysate of EH-3-248-1 expressing OxyS (+OxyS) or the control strain

EH-3-248-8 expressing no hydroxylase (−OxyS) was placed overnight

in Tris buﬀer (100.0 mM, pH 7.45) containing anhydrotetracycline

more complex derivative cofactor F₄₂₀

-Hydroxylation of Anhydrotetracycline in Saccha-

(

1, 5.4 mM), glucose (27.8 mM), NADPH (3.0 mM), and

romyces cerevisiae. The ﬁrst step required to convert

anhydrotetracyclines to tetracyclines is 6-hydroxylation. In

the biosynthesis of oxytetracycline, this step is catalyzed by

mercaptoethanol (18.5 mM). After overnight incubation, methanol

was added, the contents were mixed, and the reaction was ﬁltered

prior to MS analysis. m/z calcd for protonated anhydrotetracycline

OxyS. To test the capacity of S. cerevisiae to hydroxylate

anhydrotetracycline, we chose OxyS and anhydrotetracycline

as the enzyme−substrate pair (Scheme 1). We made this

choice because OxyS functionally expresses in Escherichia coli

(

C H N O ): 427.15. Found: 427.3 [M + H] . m/z calcd for

22 23 2 7

protonated 5a(11a)-dehydrotetracycline and for protonated 5(5a)-

dehydrotetracycline (2a and 2b, respectively, C H N O ): 443.15.

Found: 443.3 [M + H] . The expected [M + H] value for 2 holds for

either 2a or 2b as the two are isomers (Scheme 1).

and because anhydrotetracycline is commercially available.

The expression plasmid pSP-G1 was chosen to facilitate the

coexpression of a dehydrotetracycline reductase enzyme, to

append antibody tags to both the hydroxylase and the

Following this cell lysate result, the hydroxylation of

anhydrotetracycline by S. cerevisiae cells expressing OxyS was

tested by incubation of whole cells with anhydrotetracycline.

Cultures of OxyS expressing cells and control cells were

pelleted, and the pellets were resuspended and incubated

overnight with anhydrotetracycline in Tris buﬀer (pH 7.45)

prior to spectroscopic measurements. Indeed, the molecular

ion corresponding to dehydrotetracycline had over 10 times

higher ion counts in the OxyS expressing strain relative to the

control strain (Figure S1).

reductase and to constitutively express both enzymes. OxyS

was cloned into pSP-G1 under the transcriptional control of

the strong constitutive promoter TEF1 with a FLAG antibody

tag at its C-terminus.

First, we tested for the catalysis of anhydrotetracycline

hydroxylation by OxyS in S. cerevisiae cell lysate by mass

spectrometry. For the cell lysate experiment, we supplied

OxyS cells or the control −OxyS cells with the anhydrote-

tracycline starting material as well as with NADPH, a cofactor

of OxyS.

Next, the hypothesis that anhydrotetracycline is converted to

the hydroxylation intermediate, 5a(11a)-dehydrotetracycline

(2a), was tested by a larger scale reaction, puriﬁcation, and

NMR analysis of the isolates. A cell lysate of the +OxyS strain

was incubated overnight with anhydrotetracycline, and the

hydroxylation product was isolated by liquid−liquid extraction

using ethyl acetate and water followed by reverse phase HPLC

puriﬁcation. To prevent acidic and thermal degradation of the

hydroxylation intermediate, reverse phase HPLC was per-

formed with a mobile phase gradient of acetonitrile in Tris

buﬀer (pH 7.45), and NMR analysis was performed at 273 ± 5

K in methanol-d . Gradients such as acetonitrile in NH OAc,

2,27

Indeed, when a lysate of S. cerevisiae cells

expressing OxyS is incubated with anhydrotetracycline, the

molecular ion corresponding to dehydrotetracycline (2, [M +

H] = 443.3) has over 20 times higher counts compared to

when a lysate of S. cerevisiae cells not expressing OxyS is

incubated with anhydrotetracycline. In addition, the ion counts

for the molecular ions corresponding to anhydrotetracycline

(

1, [M + H] = 427.3) are decreased in the case of lysates of

cells expressing OxyS, compared to lysates of cells not

expressing OxyS, supporting the biosynthesis of dehydrote-

tracycline from anhydrotetracycline (2, Figure 1).

ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

Articles

H O:TFA, and H O, as well as ambient temperature NMR

Table 1. H-NMR Data for 5(5a)-Dehydrotetracycline (2b)

after the use of these aqueous phases led to degradation of the

intermediate (data not shown).

Notably, after the reaction of the +OxyS cell lysate with

anhydrotetracycline, 5(5a)-dehydrotetracycline (2b) was

obtained instead of the dehydrotetracycline isomer that was

anticipated, 5a(11a)-dehydrotetracycline (2a , Figure 2,

and for Tetracycline (Methanol-d , 500 MHz)

(5a)-dehydrotetracycline

no.

(2b) δ (J in Hz)

tetracycline (3) δ_H(J in Hz)

4.08, d (3.0)

2.95, ddd (12.7, 3.0, 2.5)

3.03, s

3.95, d (11.3)

3.03, dd (11.5, 6.2)

2.77, s

NMe2

5.66, d (6.2)

2.22 dddd (13.4, 3.0, 3.0, 2.5); 1.93

ddd (13.4, 13.3, 11.2)

-Me

3.05, m

1.63, s

7.17, dd (7.8, 0.7)

7.54, dd (8.1, 8.0)

6.94, dd (8.4, 0.7)

1.51, s

7.13, d (7.8)

7.28, dd (8.0, 7.9)

6.63, d (8.1)

Figure 2. Key interactions of H− H COSY and HMBC in the NMR

of 5(5a)-dehydrotetracycline (2b) and tetracycline puriﬁed from cell

lysate reaction of +OxyS and +OxyS +CtcM +FNO strains,

dehydrotetracycline isomer (2b), and not the 5a(11a)-

dehydrotetracycline form (2a), was observed in the H and

C NMR spectra, implying that 5(5a)-dehydrotetracycline

respectively (methanol-d , 500 MHz).

(

2b) is the prevalent isomer (Scheme 1). The H and

assignments of 5(5a)-dehydrotetracycline (2b) are supported

by one- and two-dimensional NMR experiments (Figure 2 and

Figures S5 −S11 ).

Scheme 2. Biosynthesis of 5(5a)-Dehydrotetracycline (2b)

and Tetracycline (3) in S. cerevisiae Cell Lysates

Reduction of the Hydroxylation Intermediate Using

Saccharomyces cerevisiae. The second step in the biosyn-

thesis of tetracycline from anhydrotetracycline is the reduction

of the 5a(11a) α,β-unsaturated double bond (Scheme 1). The

reduction at the 5a(11a) bond is known to be essential to the

antibiotic activity of the tetracyclines. For example, 7-

chlorotetracycline is over 20 times more potent than 7-

chloro-5a(11a)-dehydrotetracycline against Staphylococcus aur-

eus. To execute the reduction step, we placed OxyR, the

reductase of 5a(11a)-dehydrooxytetracycline (2a) from the

oxytetracycline pathway, under the control of the PGK1

promoter in the pSP-G1-OxyS plasmid.

The catalytic activity of OxyR is known to be dependent on

^c^o^f^a^c^t^o^r^F420, a unique cofactor not native to S. cerevisiae. F is

an intermediate in cofactor F₄₂₀biosynthesis and is known to

(5a)-dehydrotetracycline (2b) and tetracycline (3) were isolated in

this study in 25% and 24% yield, following incubation of

anhydrotetracycline (1) with lysates of +OxyS and +OxyS +CtcM

supplemented to the lysates as indicated. 2b and 3 were puriﬁed by

liquid−liquid extraction and reverse phase HPLC puriﬁcation and

characterized by NMR and HRMS (Figures S5 −S13 ). Gray squares

emphasize the carbons at which key chemical transformations occur.

F_o= 7,8-didemethyl-8-hydroxy-5-deazariboﬂavin; FNO = F₄₂₀-NADP

oxidoreductase from A. fulgidus; G6P = glucose-6-phosphate

successfully replace cofactor F₄₂₀as a substrate of F

420

24,29

reductase enzymes with similar K_mand k_c_a_tvalues.

For

FNO S. cerevisiae cells, respectively. NADPH, G6P, and F were

^e^x^a^m^p^l^e^,^a^c^o^f^a^c^t^o^r^F420 reductase from Methanobacterium

thermoautotrophicum had a K value of 19 μM with F and 34

420

μM with F_o. In another example, a cofactor F₄₂₀reductase

from Methanococcus vannielii catalyzed the reduction of F

and F with k /K values of 158 and 56 min⁻μM ,

−1

cat

respectively. To test if OxyR could be functional in S.

cerevisiae with F as a cofactor, we exogenously added synthetic

Scheme 2). All protons of 5(5a)-dehydrotetracycline (2b)

have chemical shifts within 0.3 ppm of the corresponding

protons in tetracycline except for the protons attached to the

C and C positions (Table 1). As expected, a proton on C

F to a lysate of S. cerevisiae cells expressing OxyS, OxyR, and

an F reductase. We employed F in a concentration of 400

μM, about 10× the experimental K_mvalues of F with F

420

reductases. We tested if F₄₂₀reductases can function as F_o

reductases in our system. Toward this end, we used F₄₂₀

reductases from three hosts, M. tuberculosis, A. fulgidus, and S.

griseus, by expression under the control of pGPD (pTDH3) on

exists in tetracycline but not in 5(5a)-dehydrotetracycline

2b). In addition, while tetracycline has two aliphatic protons

(

on C at δ 1.93 and 2.22, 5(5a)-dehydrotetracycline (2b) has

32−35

only one C proton at δ 5.66, typical of oleﬁnic protons. By

pRS413.

The reaction mixture containing the S. cerevisiae

contrast, 5a(11a)-dehydrotetracycline (2a), the expected

product of OxyS hydroxylation of anhydrotetracycline, should

cell lysate in Tris buﬀer (pH 7.45) also contained

anhydrotetracycline, glucose, and NADPH. In the reactions

with F₄₂₀reductase from M. tuberculosis, glucose-6-phosphate

(G6P), the reducing agent used by this enzyme, was also

included for F reduction. Since the other two F reductases

have two aliphatic protons on C and no additional oleﬁnic

proton (Scheme 1). The interconversion between 5(5a)-

dehydrotetracycline (2b) and 5a(11a)-dehydrotetracycline

420

(

2a) is supported by a D−H exchange experiment, where

use NADPH as the reducing agent of F₄₂₀, and given that

NADPH was already included in the reaction setup as a

cofactor for OxyS, no additional reducing agent was added to

the 5-H peak is eliminated over time at 300 K (Figure S2) in

methanol-d . Despite this interconversion, only the 5(5a)-

ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

Articles

the reactions of F₄₂₀reductases from A. fulgidus and S. griseus.

We found that the levels of the molecular ion peak

corresponding to tetracycline ([M + H] = 445.2), the

reduction product of 5a(11a)-dehydrotetracycline (2a, [M +

H] = 443.2), increased when G6P was added (Figure 3a vs c,

dashed line, Figure S3). Contrary to the expectation that G6P

serves as a substrate for the F reductase, we found that neither

F , the F reductase, nor OxyR contributed to the increase in

the levels of the molecular ion peak corresponding to

tetracycline in the presence of G6P (data not shown).

However, the low conversion rates of this setup did not

allow the puriﬁcation of tetracycline in levels permitting NMR

analysis, and we therefore turned to test alternative reductase

enzymes.

Given that OxyS performs one hydroxylation in S. cerevisiae

Figure 1) as opposed to two in vitro and in S. rimosus, it was

(

logical to test an alternative dehydrotetracycline reductase with

OxyR. The native substrate for OxyR is hypothesized to be the

doubly hydroxylated 5a(11a)-dehydrooxytetracycline (4) and

not the singly hydroxylated 5a(11a)-dehydrotetracycline (2a,

Scheme 1). Furthermore, using the alternative enzyme CtcM

from the chlortetracycline pathway instead of OxyR yielded in

vitro an increased ratio of tetracycline to oxytetracycline.

Another reasonable alternative candidate to OxyR was DacO4

from the dactylocycline pathway, an additional OxyR

homologue, whose hypothetical native substrate, 5a(11a)-

dehydrodactylocyclinone, is also not hydroxylated at C₅. We

therefore tested two candidate dehydrotetracycline reductases,

CtcM and DacO4, along with synthetic F and three F

420

reductase candidates, from M. tuberculosis, A. fulgidus, or S.

14,23,25,32,34,35

griseus, leading to six combinations.

Gratifyingly,

incubating the cell lysate of the strain encoding OxyS, CtcM,

and FNO from A. fulgidus with anhydrotetracycline resulted in

an F -dependent peak corresponding to tetracycline (Figure 3b

vs a, solid line). As expected, such an F -dependent increase in

the peak corresponding to tetracycline was not observed in the

control strain lacking CtcM and FNO (Figure 3b vs a, dotted

line).

In light of the G6P-dependent increase of the molecular ion

counts corresponding to tetracycline even in the absence of

CtcM (Figure S3 ), a potential synergy between F and G6P

was tested. The cell lysate of the +OxyS +CtcM +FNO strain

was incubated with anhydrotetracycline and NADPH in the

presence of both F and G6P in Tris buﬀer (pH 7.45). Indeed,

a major decrease in the molecular ion counts corresponding to

dehydrotetracycline (2) and a major increase in the molecular

ion counts corresponding to tetracycline (3) were observed

(

Figure 3d vs a−c, solid line). As expected, this result was not

observed in the +OxyS −CtcM −FNO cell lysate (Figure 3d vs

a−c, dotted line).

The conversion of anhydrotetracycline to tetracycline by the

OxyS +CtcM +FNO cell lysate in the presence of F and G6P

was conﬁrmed by puriﬁcation and NMR characterization

Scheme 2). Tetracycline was isolated in 24% yield using

(

liquid−liquid extraction with ethyl acetate and water followed

by reverse-phase HPLC separation using a mobile phase

gradient of acetonitrile in H O/TFA (99.9:0.1). Tetracycline

Figure 3. Mass spectrometry analysis of anhydrotetracycline

hydroxylation and reduction in lysates of S. cerevisiae cells expressing

was obtained from the cell lysates at a yield of 36 mg/L of the

original culture of the cells. The identity of the tetracycline

OxyS, CtcM, and FNO in the presence of G6P and F . Cell lysates of

thus obtained to a tetracycline standard was conﬁrmed by H

OxyS +CtcM +FNO strain EH-6-77-3 () or the +OxyS −CtcM

NMR and HRMS (Figures S12 and S13 ). In addition, the

conversion of anhydrotetracycline to tetracycline upon

−

FNO control strain EH-3-204-9 (···) were placed overnight in Tris

buﬀer (100.0 mM, pH 7.45) containing anhydrotetracycline (5.4

incubation of anhydrotetracycline and F with unlyzed yeast

mM), glucose (27.8 mM), NADPH (3.0 mM), F (0.4 mM (+FO) or

ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

Articles

Figure 3. continued

for our ability to purify and analyze 5(5a)-dehydrotetracycline

2b) from the +OxyS S. cerevsiae cell lysate. First, the

(

additional degradation pathways that become possible in the

presence of the additional 5-hydroxy installed by OxyS in vitro

and in Streptomyces but not in our study. For example,

anhydrooxytetracycline (5-hydroxyanhydrotetracycline) is

known to undergo B-ring scission degradation, which is not

mM (−FO)), glucose-6-phosphate (10.0 mM (+G6P) or 0 mM

(

−G6P)), and mercaptoethanol (18.5 mM). m/z calcd for protonated

anhydrotetracycline (1, C H N O ): 427.15. Found, 427.0−427.2

[

M + H] . m/z calcd for protonated 5a(11a)-dehydrotetracycline or

(5a)-dehydrotetracycline (2a and 2b, respectively, C H N O ):

43.15. Found: 443.1−443.2 [M + H] . m/z calcd for protonated

22 23 2 8

possible in anhydrotetracycline. Second, in our study, the

tetracycline (3, C H N O ): 445.16. Found: 445.1−445.2 [M +

H] . The expected [M + H] value for 2 holds for either 2a or 2b as

hydroxylated product of anhydrotetracycline was protected

from acid, while in the in vitro study of OxyS, acidic organic

the two are isomers (Scheme 1). F = 7,8-didemethyl-8-hydroxy-5-

deazariboﬂavin; FNO = F₄₂₀-NADP oxidoreductase from A. fulgidus;

G6P = glucose-6-phosphate.

extraction was employed. Anhydrooxytetracycline is known

to undergo B-ring scission speciﬁcally in the presence of dilute

acid.

Our characterization of 5(5a)-dehydrotetracycline (2b)

enhances the current understanding of the last steps of

chlortetracycline and oxytetracycline biosynthesis and the

diﬀerences between these pathways. First, prior to our study,

5(5a)-dehydrotetracycline (2b) was an uncharacterized hypo-

thetical intermediate in the OxyS hydroxylation of anhydrote-

tracycline. By characterizing 5(5a)-dehydrotetracycline (2b),

our study supports the hypothesized mechanism of OxyS

hydroxylation in the biosynthetic pathway of oxytetracycline

cells expressing OxyS, CtcM, and FNO in Tris buﬀer is

supported by mass spectrometry (Figure S4).

DISCUSSION

■

This study demonstrates the use of S. cerevisiae for the ﬁnal

steps of tetracycline biosynthesis, the hydroxylation of

anhydrotetracycline, and the reduction of dehydrotetracycline

(

Scheme 1). These steps are key toward the total biosynthesis

(

5, Scheme 1, Figure 2). Second, prior to our study, it was

of tetracyclines in S. cerevisiae in an eﬀort to biosynthesize new

tetracycline analogs to combat antibiotic resistance. Heterol-

ogous biosynthesis of oxytetracycline was shown in Strepto-

myces lividans K4−11, a much closer relative of the original

hypothesized that the double hydroxylation performed by

OxyS as opposed to the single hydroxylation performed by

CtcN results from structural diﬀerences between OxyS and

CtcN. An increased stability of 5(5a)-dehydrotetracycline

2b) over 5a(11a)-dehydrotetracycline (2a), supported by our

oxytetracycline biosynthesis host, Streptomyces rimosus. The

choice of S. cerevisiae as a heterologous host in our study

enabled the biosynthesis of tetracycline, instead of oxy-

tetracycline. This diﬀerence resulted from two reasons: ﬁrst,

a single hydroxylation catalyzed by OxyS from the oxy-

tetracycline pathway in S. cerevisiae, as opposed to a double

hydroxylation in Streptomyces and in vitro (Figure 1), and

second, our use of CtcM from the chlortetracycline pathway,

instead of OxyR from the oxytetracycline pathway, as the

dehydrotetracycline reductase (Figure 3). Possible reasons for

the single hydroxylation of OxyS in S. cerevisiae instead of the

double hydroxylation observed in puriﬁed enzymes could be

diﬀerences in the post-translational modiﬁcations of OxyS or

diﬀerences in the chemical environment of OxyS, such as pH

diﬀerences. Future research could determine the relevant factor

for this diﬀerence.

(

study (Scheme 1, Figure 2), can promote an additional

hypothesis. Namely, that structural diﬀerences between the

dehydrotetracycline intermediates of the two pathways can also

contribute to the diﬀerence in the number of hydroxylations

between the two pathways. Speciﬁcally, 5(5a)-dehydrotetracy-

cline (2b) is the substrate for a second hydroxylation step, and

hence its increased stability over 5a(11a)-dehydrotetracycline

(2a), the substrate for reduction, can promote a second

hydroxylation (Scheme 1). Third, among the dehydrotetracy-

cline intermediates in the biosynthesis of chlortetracycline,

oxytetracycline, and tetracycline, the only dehydrotetracycline

previously characterized is 5a(11a)-dehydrochlortetracycline

28,38

(2a, Figure S14).

Future studies can determine whether

the lack of the 7-chloro substituent can be a contributing factor

to an increased stability of 5(5a)-dehydrotetracycline (2b)

over 5a(11a)-dehydrotetracycline (2a).

Production of tetracycline in our work requires supplemen-

tation of anhydrotetracycline to S. cerevisiae cells or cell lysates.

For a total biosynthesis of tetracycline and its analogs in S.

cerevisiae, the additional enzymes coding for tetracycline

biosynthesis would have to be heterologously expressed as

well. Our work solves important challenges toward the goal of

total biosynthesis of tetracyclines in S. cerevisiae, such as the

challenge of reducing dehydrotetracycline to tetracycline in a

heterologous host that does not biosynthesize the native

cofactor, coenzyme F₄₂₀. Importantly, for industrial production

of tetracyclines by S. cerevisiae, the yield of tetracycline would

need to be improved from our current yield of 36 mg/L to

yields that are closer to yields of tetracyclines from industrial

strains, such as the industrial strains of Streptomyes rimosus

Furthermore, we have shown that F , a synthetically

accessible precursor in the biosynthesis of cofactor F420, can

successfully replace F420 in the biosynthesis of tetracycline from

anhydrotetracycline (Figure 3). To function eﬀectively in our

setup, F

reductase FNO and by the dehydrotetracycline reductase

CtcM. The ability of F to replace F420 in enzymatic reactions

of F₂420 reductases was previously known in the litera-

needed to be accepted as a substrate both by the F

4,29−31,39

ture.

However, our study critically shows that this

ability further extends to F₄₂₀-dependent biosynthetic enzymes

such as CtcM from the chlortetracycline pathway. This result is

important since F₄₂₀is a unique cofactor that is not native to

producing oxytetracycline at 100 g/L.

common heterologous hosts such as S. cerevisiae and E. coli.

Our study is the ﬁrst to isolate and analyze a dehydrote-

tracycline intermediate when OxyS is expressed in the absence

of a dehydrotetracycline reductase, such as OxyR or CtcM

Previous attempts at biosynthesis of oxytetracycline using

puriﬁed enzymes with exogenously added F₄₂₀or using the

more closely related heterologous host Streptomyces lividans

that natively biosynthesizes F₄₂₀have been employing cofactor

F₄₂₀and were thus limited in the diversity of possible

heterologous hosts that can be used or in the scale of

(Figures 1 and 2). The reaction product of OxyS rapidly

degraded both in vitro, when OxyR was not concurrently used,

and in vivo, in ΔoxyR S. lividans. Two explanations are noted

ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

Articles

tetracycline production, respectively. Moreover, future studies

MATERIALS AND METHODS

■

can now incorporate the biosynthesis of F into a tetracycline

General Methods. Absorption and ﬂuorescence spectra were

recorded on an Inﬁnite-M200 ﬂuorescent spectrometer. DNA

sequences were purchased from IDT. Polymerases, restriction

enzymes, and Gibson Assembly mix were purchased from New

England Biolabs. Sanger sequencing was performed by Genewiz. Yeast

strains were grown at 30 °C, and shaker settings were 200 rpm, unless

otherwise indicated. Yeast transformations were done using the

^bⁱ^o^s^yⁿ^t^h^e^sⁱ^s^s^y^s^t^e^mⁱⁿ^S^.^c^e^r^e^vⁱ^sⁱ^a^e^sⁱⁿ^c^e^Fo is only one

biosynthetic step removed from common S. cerevisiae

metabolites, tyrosine and 5-amino-6-ribitylamino-2,4-

40−43

(

1H,3H)-pyrimidinedione.

Our approach of using F to

replace F₄₂₀in the heterologous biosynthesis of tetracyclines

could be further extended to other F₄₂₀-dependent pathways,

such as the pathways leading to lincosamides and aminoglyco-

lithium acetate method. Plasmids were cloned and ampliﬁed using

Gibson Assembly and cloning strain C3040 (New England Biolabs).

Unless otherwise indicated, yeast strains were grown on synthetic

minimal media lacking histidine and/or uracil and/or tryptophan

and/or leucine, as indicated by the abbreviation HUTL. Yeast strain

sides.

The synergistic eﬀect of F and G6P could be explained by a

G6P-dependent increase of NADPH pools from NADP ,

catalyzed by G6P dehydrogenase, leading to the reduction of

patches were obtained from glycerol stocks by streaking on an agar

plate of synthetic medium lacking the appropriate amino acid

markers, incubating at 30 °C for 3 days, patching single colonies onto

a fresh agar plate, and incubating at 30 °C overnight. Protein

homology was calculated by BLAST (https://blast.ncbi.nlm.nih.gov)

using the standard sequence alignment parameters. DataExpress was

used to analyze Advion CMS data, and MassLynx was used to analyze

Waters XEVO QTOF data.

Codon optimization by COOL (http://cool.syncti.org/index.php)

was used for all hydroxylases and reductases unless noted explicitly

that JCAT or no optimization was used (http://www.jcat.de/).

Optimization parameters chosen were individual codon use, a codon

context GC content of 39.3%, and the S. cerevisiae organism. The

following restriction sites were generally excluded: GAGACC,

GGTCTC, GGATCC, GAGCTC, CTCGAG, GAAGAC, GTCTTC,

CGTCTC, GAGACG, GAATTC, TTAATTAA, TCTAGA, AC-

TAGT, CCCGGG, CTGCAG, AAGCTT, GTCGAC, ACGCGT,

GGTACC, GCGGCCGC, AGATCT, GGCCGGCC, CCGCGG,

GCTAGC, and CCATGG.

F to F H (Scheme S1 ). Interestingly, G6P increases the ion

counts corresponding to protonated tetracycline also in the

absence of F , albeit to a lesser degree than in the presence of

F_o(Figure 3). Future research can determine whether a native

S. cerevisiae enzyme is catalyzing the reduction of 5a(11a)-

dehydrotetracycline (2a) in the absence of F in a G6P-

dependent manner. Such G6P dependence could be direct or

through an increase in another cellular reducing agent, such as

NADPH.

Beyond the use of S. cerevisiae toward tetracycline biosyn-

thesis, S. cerevisiae can oﬀer access to novel tetracycline

analogs. This opportunity exists because of the widely available

tools for genetically modifying S. cerevisiae as well as enhanced

45−48

accessibility of biosynthetic enzymes such as P450s.

Speciﬁcally, S. cerevisiae can be used to express alternative

hydroxylases such as DacO1 and CtcN that proved previously

to be insoluble when expressed in other heterologous hosts

Preparative HPLC was carried out with a C-18 5 μ column, 250 ×

0 mm, eluent given in parentheses. NMR spectra were obtained

such as E. coli and Streptomyces. Such enzymes can

hydroxylate alternative anhydrotetracycline substrates, as well

as lead to hydroxylated products of the opposite stereo-

chemistry in the 6-position, thereby covering additional

chemical space. Importantly, such chemical space is not

covered by existing methods of synthesizing tetracyclines,

despite the promise of 6-position tetracycline analogs for

using Bruker 400 or 500 MHz instruments, as indicated. Unless stated

otherwise, mass spectroscopy measurements were performed on an

Advion CMS mass spectrometer equipped with an atmospheric

pressure chemical ionization (APCI) source. HRMS spectra and MS

analyses of whole cell supernatants were taken on a Waters

ACQUITY UPLC XEVO QTOF equipped with a BEH C18 column

(2.1 × 50 mm) at 30 °C with a ﬂow rate of 0.8 mL/min. Unless stated

otherwise, all reagents, salts, and solvents were purchased from

commercial sources and used without further puriﬁcation.

potent antibiotic activity. Moreover, the use of S. cerevisiae

for the conversion of anhydrotetracyclines to tetracyclines can

readily utilize fungal anhydrotetracycline analogs for further

17,50,51

General Protocol for Hydroxylation/Reduction Assay with

Cell Lysate. Fresh patches of strains harboring the plasmid for the

hydroxylation with/without the reduction enzyme and with/without

the plasmid for the F420 reductase enzyme and control strains were

diversity generation.

CONCLUSION

■

−

This study shows that with expression of three heterologous

genes and the exogenous supply of a cofactor intermediate, S.

cerevisiae could be used for the ﬁnal steps of tetracycline

biosynthesis. This study further conﬁrms the structure of a

hypothetical intermediate in the biosynthesis of oxytetracy-

cline, providing a potential explanation for the structural

diﬀerence between the chlortetracycline and oxytetracycline

natural products. Additionally, of signiﬁcant practical impor-

inoculated in 5 mL of selective media (U or HU ) in 15 mL culture

^t^u^b^e^s⁽^C^o^rⁿⁱⁿ^g³⁵²⁰⁵⁹⁾^aⁿ^d^p^l^a^c^e^dⁱⁿ^a^s^h^a^k^e^r^o^v^e^rⁿⁱ^g^h^t^t^o^O^D600

−3. Overnight cultures were used to inoculate 100 mL of selective

− −

media (U or HU ) cultures in 500 mL conical ﬂasks with a starting

OD₆₀₀of 0.01−0.05. Cells were grown to ﬁnal OD of 0.6−0.8

600

before pelleting in two 50 mL tubes (Corning 352098) at 4 °C and

000 rpm for 20 min. Each pellet was redissolved in 0.5 mL of H O,

and the suspension was distributed into two presterilized 1.5 mL

Eppendorf tubes and pelleted at 14 000 rpm for 10 min at 4 °C.

Pellets were stored at −20 °C prior to further use.

tance, the simple cofactor F is shown to substitute eﬃciently

for cofactor F₄₂₀

Pellets were weighed and thawed on ice. A 99:1 mixture of Y-PER

yeast protein extraction reagent (ThermoFisher Scientiﬁc 78991) and

HALT protease inhibitor cocktail (ThermoFisher Scientiﬁc PI87786)

was added in a ratio of 3 μL of mixture per milligram pellet and placed

on an orbital shaker for 20 min at 22 °C, followed by 10 min of

centrifugation at 14 000 rpm at 4 °C, and the cell lysate was

transferred to a new 1.5 mL Eppendorf tube, kept on ice, and used

within 1 h.

While at an early stage of development, the results presented

here are the ﬁrst step toward the total biosynthesis of

tetracycline and its analogs in S. cerevisiae. These tetracycline

analogs could have the potential to combat existing and future

mechanisms of bacterial antibiotic resistance. Furthermore,

they could introduce unique chemical handles for further

derivatization by semisynthesis. Finally, in the future, yeast

secreting tetracyclines could have myriad applications not yet

envisioned signiﬁcantly impacting the ﬁeld of synthetic

biology.

The cell lysate (0.080 mL) was added as the last component to a 4

mL vial (Chemglass CG-4900−01) containing 0.280 mL of 143.0

mM Tris (pH 7.45), 7.7 mM anhydrotetracycline·HCl (AdipoGen

CDX-A0197-M500), 4.3 mM NADPH tetrasodium hydrate (Sigma-

ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

The Supporting Information is available free of charge at

Articles

Aldrich N7505), 26.4 mM mercaptoehtanol, 0.5 mM of F_o(in

experiments labeled + FO, 0 mM in all other experiments), 14.3 mM

glucose-6-phosphate (in experiments labeled + G6P, 0 mM glucose-6-

phosphate in all other experiments), and 0.040 mL of glucose (278.0

mM) for ﬁnal concentrations of 100.0 mM Tris, 5.4 mM

anhydrotetracycline·HCl, 3 mM NADPH, 18.5 mM mercaptoethanol,

ﬁnal concentrations of 100.0 mM Tris, 5.4 mM anhydrotetracycline·

HCl, 3.0 mM NADPH, 18.5 mM mercaptoethanol, and 27.8 mM

glucose. A septum was placed on top of the ﬂask through which a

needle was inserted to allow air exchange, and the reaction was left at

22 °C overnight. After 14 h, the aqueous mixture was extracted two

times with EtOAc. The combined organic fraction was extracted with

water. MeCN was then added to the combined aqueous phase, and it

was then dried at 20 °C. The contents were then dissolved in a

mixture of water, MeCN, and MeOH and puriﬁed by preparative RP-

HPLC (1−20% MeCN in Tris (pH 7.45, 20.0 mM, 60 min)) to

or 0.4 mM F as indicated, 0 or 10.0 mM G6P as indicated, and 27.8

mM glucose. A septum was placed on top of the vial through which a

needle was inserted to allow air exchange and the reaction was left at

2 °C overnight. After the night, 1 mL of MeOH was added. The

contents were mixed, and the reaction was ﬁltered through a PTFE

.2 μm ﬁlter (Acrodisc 4423) prior to analysis by mass and UV/vis

aﬀord 5(5a)-dehydrotetracycline (2b) after lyophilization (11.1 mg,

25% yield) as a yellow solid. 2b: H NMR (500 MHz, methanol-d ):

spectrometry.

δ 7.28 (dd, J = 8.0, 7.9 Hz, 1 H), 7.13 (d, J = 7.8 Hz, 1 H), 6.63 (d, J =

8.1 Hz, 1 H), 5.66 (d, J = 6.2 Hz, 1 H), 3.95 (d, J = 11.3 Hz, 1 H),

Protocol for Hydroxylation/Reduction Assay with Whole

Cells. Fresh patches of strains harboring the plasmid for the

hydroxylation and/or reduction enzyme and control strains were

inoculated in 5 mL of selective media (U or HU ) in 15 mL culture

tubes (Corning 352059) and placed in shaker overnight to OD₆₀₀2−

3.03 (dd, J = 11.5, 6.2 Hz, 1 H), 2.77 (s, 6 H), 1.51 (s, 3 H).

NMR (500 MHz, D O): δ 192.6, 188.3, 182.9, 180.4, 172.1, 160.7,

−

148.1, 146.3, 134.3, 116.3, 116.1,115.0, 106.1, 103.5, 102.7, 77.8, 73.1,

69.1, 42.4, 39.3, 34.4. HRMS (ES+), m/z calcd for C H N O :

−

. Overnight cultures were used to inoculate 100 mL of selective

443.1449. Found: 443.1415 [M + H] . HRMS (ES ), m/z calcd for

−

media (U or HU ) cultures in 500 mL conical ﬂasks with a starting

OD₆₀₀of 0.05−0.1. Cells were grown for 22 h before being placed at

C H N O : 441.1303. Found: 441.1318 [M − H] .

Biosynthesis of Tetracycline Using S. cerevisiae. Fresh

5 °C for an additional 10−12 h. Cells were then pelleted in two 50

patches of EH-6-77-3 harboring the plasmids encoding OxyS,

−

mL tubes (Corning 352098) at 10 °C and 3500 rpm for 5 min. Each

pellet was redissolved in 0.5 mL of H O, and the suspension was

CtcM, and FNO were inoculated in selective media (HU ) in 15

mL culture tubes (Corning 352059) and placed in a shaker overnight

distributed into a presterilized 1.5 mL Eppendorf tube and pelleted at

to OD₆₀₀2−3. Overnight cultures were used to inoculate 100 mL of

−

1 000 rpm for 3 min at 10 °C. Pellets were placed on ice and used

selective media (HU ) cultures in 500 mL conical ﬂasks with a

within 1 h.

starting OD₆₀₀of 0.01−0.05. Cells were grown to a ﬁnal OD of 0.75

600

When strains EH-3-98-6 and EH-3-80-3 were used, pellets from 50

mL of culture were redissolved in H O (1.025 mL) and added as the

last component to 15 mL culture tubes (Corning 352059) containing

before pelleting in 500 mL tubes at 4 °C and 6000 rpm. The pellet

was redissolved in 25 mL of H O, and the suspension was distributed

into 50 mL falcon tubes and pelleted at 4 °C and 4000 rpm. The

pellet was then transferred into four presterilized 1.5 mL Eppendorf

tubes and pelleted at 14 000 rpm for 10 min at 4 °C. Pellets were

stored at −20 °C prior to further use.

Pellets from 50 mL of culture were weighed and thawed on ice. A

99:1 mixture of Y-PER yeast protein extraction reagent (Thermo-

Fisher Scientiﬁc 78991) and HALT protease inhibitor cocktail

(ThermoFisher Scientiﬁc PI87786) was added in a ratio of 3 μL of

mixture per milligram of pellet and placed on an orbital shaker for 20

min at 22 °C, followed by 10 min of centrifugation at 14 000 rpm at 4

°C, and the cell lysate was transferred to a new 1.5 mL Eppendorf

tube, kept on ice, and used within 1 h.

−

.100 mL of 8 mg mL anhydrotetracycline·HCl, 0.125 mL of

glucose solution in H O (40%), and 0.250 mL of 1 M Tris buﬀer at

pH 7.45 and were placed in a shaker at 350 rpm at 21 °C for 27 h.

Cultures were then pelleted, and the supernatant was diluted into

H O before being used for mass and UV/vis spectroscopy.

When strains EH-6-77-3 and EH-3-204-9 were used, pelleted

unlyzed cells were redissolved in H O (0.4 mL) and added as the last

component to 15 mL culture tubes (Corning 352059) containing

−1

.100 mL of 8 mg mL anhydrotetracycline·HCl, 0.125 mL of

glucose solution in H O (40%), 0.250 mL of 1 M Tris buﬀer at pH

.45, and 0.625 mL of 1.5 mM F (+FO) or 0.625 mL of H O (−FO)

and were placed in a shaker at 350 rpm at 21 °C for 27 h. Cultures

were then pelleted, and the supernatant was diluted into H O before

The cell lysate (0.440 mL) was added as the last component to four

borosilicate vials of 4 mL each (4 × 0.110 mL) containing 1.760 mL

being used for mass and UV/vis spectroscopy.

Biosynthesis of (5,5a)-Dehydrotetracycline (2b) Using S.

cerevisiae. Fresh patches of EH-3-248-1 harboring the plasmid

(

0.440 mL each) of 125.1 mM Tris (pH 7.45), 6.7 mM

anhydrotetracycline·HCl (AdipoGen CDX-A0197-M500), 3.8 mM

NADPH tetrasodium hydrate (Sigma-Aldrich N7505), 23.1 mM

−

encoding OxyS were inoculated in 2 × 5 mL selective media (U ) in

mercaptoethanol, 0.5 mM F , 125.1 mM G6P, and 34.7 mM glucose

5 mL culture tubes (Corning 352059) and placed in a shaker

for ﬁnal concentrations of 100.0 mM Tris, 5.4 mM anhydrotetracy-

cline·HCl, 3.0 mM NADPH, 18.5 mM mercaptoethanol, 0.4 mM F_o,

^o^v^e^rⁿⁱ^g^h^t^t^o^O^D⁶00 2−3. Overnight cultures were used to inoculate

−

00 mL of selective media (U ) cultures in 2 L conical ﬂasks with a

00.0 mM G6P, and 27.8 mM glucose. A septum was placed on top of

^s^t^a^r^tⁱⁿ^g^O^D⁶00 of 0.01−0.05. Cells were grown to a ﬁnal OD of 0.75

600

each vial through which a needle was inserted to allow air exchange,

and the reaction was left at 22 °C overnight. After 16 h, the aqueous

mixture was extracted two times with EtOAc (7.5 mL in each round).

The combined organic fraction was extracted with water (7.5 mL).

MeOH was then added to the combined aqueous phase, and it was

then dried at 25−30 °C. The contents were then dissolved in a

mixture of water, MeCN, and MeOH and puriﬁed by preparative RP-

before pelleting in 500 mL tubes at 4 °C and 6000 rpm. The pellet

was redissolved in 25 mL of H O, and the suspension was distributed

into 50 mL falcon tubes and pelleted at 4 °C and 4000 rpm. The

pellet was then transferred into four presterilized 1.5 mL Eppendorf

tubes and pelleted at 14 000 rpm for 10 min at 4 °C. Pellets were

stored at −20 °C prior to further use.

Pellets were weighed and thawed on ice. A 99:1 mixture of Y-PER

yeast protein extraction reagent (ThermoFisher Scientiﬁc 78991) and

HALT protease inhibitor cocktail (ThermoFisher Scientiﬁc PI87786)

was added in a ratio of 3 μL of mixture per milligram pellet and placed

on an orbital shaker for 20 min at 22 °C, followed by 10 min of

centrifugation at 14 000 rpm at 4 °C, and the cell lysate was

transferred to a new 1.5 mL Eppendorf tube, kept on ice, and used

within 1 h.

The cell lysate (3.2 mL) was added as the last component to a 50

mL round-bottom ﬂask with a stir bar containing 12.8 mL of 125.1

mM Tris (pH 7.45), 6.7 mM anhydrotetracycline·HCl (AdipoGen

CDX-A0197-M500), 3.8 mM NADPH tetrasodium hydrate (Sigma-

Aldrich N7505), 23.1 mM mercaptoethanol, and 34.7 mM glucose for

HPLC (1−50% MeCN in 99.9:0.1% H O/TFA, 90 min) to aﬀord,

after drying, tetracycline (1.8 mg, 24% yield) as a yellow solid with

HRMS and H NMR spectra identical to the tetracycline standard

(

Figures S12 and S13).

ASSOCIATED CONTENT

■

sı Supporting Information

https://pubs.acs.org/doi/10.1021/acschembio.1c00259.

Supplementary ﬁgures, tables, and schemes (PDF)

ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

(9) Speer, B. S., Shoemaker, N. B., and Salyers, A. A. (1992)

Articles

Hope, W., McCarthy, J. S., Mills, J., Mouton, J. W., and Paterson, D.

AUTHOR INFORMATION

■

L., Ed.; CRC Press: Boca Raton, FL, 2010; Vol. 1, pp 851−869.

Corresponding Author

Virginia W. Cornish − Department of Chemistry, Columbia

University, New York, New York 10027, United States;

Department of Systems Biology, Columbia University, New

York, New York 10032, United States; Email: vc114@

columbia.edu

Bacterial resistance to tetracycline: mechanisms, transfer, and clinical

significance. Clin. Microbiol. Rev. 5, 387−99.

(10) Connell, S. R., Tracz, D. M., Nierhaus, K. H., and Taylor, D. E.

(2003) Ribosomal protection proteins and their mechanism of

tetracycline resistance. Antimicrob. Agents Chemother. 47, 3675−81.

(11) Gerrits, M. M., de Zoete, M. R., Arents, N. L., Kuipers, E. J.,

Authors

and Kusters, J. G. (2002) 16S rRNA mutation-mediated tetracycline

resistance in Helicobacter pylori . Antimicrob. Agents Chemother. 46,

2996−3000.

Ehud Herbst − Department of Chemistry, Columbia

University, New York, New York 10027, United States;

Present Address: Department of Molecular Genetics,

Weizmann Institute of Science, Rehovot 76100, Israel;

orcid.org/0000-0002-5570-5804

Arden Lee − Department of Chemistry, Columbia University,

New York, New York 10027, United States

Yi Tang − Department of Chemical and Biomolecular

Engineering and Department of Chemistry and Biochemistry,

University of California, Los Angeles, California 90095,

United States; orcid.org/0000-0003-1597-0141

Scott A. Snyder − Department of Chemistry, University of

Chicago, Chicago, Illinois 60637, United States;

orcid.org/0000-0003-3594-8769

(12) Speer, B. S., and Salyers, A. A. (1989) Novel aerobic

tetracycline resistance gene that chemically modifies tetracycline. J.

Bacteriol. 171, 148−53.

(13) Host’álek, Z., and Vanek, Z. Biosynthesis of the Tetracyclines.

̌ ̌

In The Tetracyclines; Hlavka, J. J., and Boothe, J. H., Eds.; Springer:

Berlin, 1985; pp 137−178.

(14) Wang, P., Kim, W., Pickens, L. B., Gao, X., and Tang, Y. (2012)

Heterologous Expression and Manipulation of Three Tetracycline

Biosynthetic Pathways. Angew. Chem., Int. Ed. 51 , 11136 −11140.

(15) Liu, F., and Myers, A. G. (2016) Development of a platform for

the discovery and practical synthesis of new tetracycline antibiotics.

Curr. Opin. Chem. Biol. 32, 48−57.

(16) Charest, M. G., Lerner, C. D., Brubaker, J. D., Siegel, D. R., and

Myers, A. G. (2005) A convergent enantioselective route to

structurally diverse 6-deoxytetracycline antibiotics. Science 308,

Complete contact information is available at:

https://pubs.acs.org/10.1021/acschembio.1c00259

(

95−8.

17) Li, Y. R., Chooi, Y. H., Sheng, Y. W., Valentine, J. S., and Tang,

Funding

Y. (2011) Comparative Characterization of Fungal Anthracenone and

Naphthacenedione Biosynthetic Pathways Reveals an alpha-Hydrox-

ylation-Dependent Claisen-like Cyclization Catalyzed by a Dimanga-

nese Thioesterase. J. Am. Chem. Soc. 133, 15773−15785.

This work was supported by National Institutes of Health

Grants R01 AI110794 and R01 GM134293.

Notes

(18) Ro, D. K., Paradise, E. M., Ouellet, M., Fisher, K. J., Newman,

The authors declare no competing ﬁnancial interest.

K. L., Ndungu, J. M., Ho, K. A., Eachus, R. A., Ham, T. S., Kirby, J.,

Chang, M. C., Withers, S. T., Shiba, Y., Sarpong, R., and Keasling, J.

D. (2006) Production of the antimalarial drug precursor artemisinic

acid in engineered yeast. Nature 440, 940−3.

ACKNOWLEDGMENTS

■

The authors thank F. W. Foss Jr. for the generous gift of

synthetic F , L. Reid for cloning plasmids LR-23-A-C3 and LR-

(19) Galanie, S., Thodey, K., Trenchard, I. J., Filsinger Interrante,

M., and Smolke, C. D. (2015) Complete biosynthesis of opioids in

3-C-C7, G. A. Ellestad for helpful discussions, and P. A.

yeast. Science 349, 1095−1100.

Baldera-Aguayo for critical reading of the manuscript.

(20) Luo, X., Reiter, M. A., d’Espaux, L., Wong, J., Denby, C. M.,

Lechner, A., Zhang, Y., Grzybowski, A. T., Harth, S., Lin, W., Lee, H.,

Yu, C., Shin, J., Deng, K., Benites, V. T., Wang, G., Baidoo, E. E. K.,

Chen, Y., Dev, I., Petzold, C. J., and Keasling, J. D. (2019) Complete

biosynthesis of cannabinoids and their unnatural analogues in yeast.

Nature 567, 123−126.

REFERENCES

■

(

1) Antibiotic Resistance Threats in the United States; U.S. Department

of Health and Human Services, Center for Disease Control and

Prevention, 2013.

2) No Time to Wait: Securing the Future from Drug-Resistant

Infections; Report to the Secretary-General of the United Nations, UN

Interagency Coordination Group on Antimicrobial Resistance, 2019.

3) Mullard, A. (2014) Momentum builds around new antibiotic

business models. Nat. Rev. Drug Discovery 13, 711 −713.

4) Brown, D. G., and Wobst, H. J. (2021) A Decade of FDA-

(

(21) Awan, A. R., Blount, B. A., Bell, D. J., Shaw, W. M., Ho, J. C. H.,

McKiernan, R. M., and Ellis, T. (2017) Biosynthesis of the antibiotic

nonribosomal peptide penicillin in baker ’s yeast. Nat. Commun. 8,

22) Wang, P., Bashiri, G., Gao, X., Sawaya, M. R., and Tang, Y.

5202.

2013) Uncovering the Enzymes that Catalyze the Final Steps in

Oxytetracycline Biosynthesis. J. Am. Chem. Soc. 135, 7138−7141.

23) Zhu, T., Cheng, X., Liu, Y., Deng, Z., and You, D. (2013)

Approved Drugs (2010 −2019): Trends and Future Directions. J. Med.

Chem. 64, 2312−2338.

Deciphering and engineering of the final step halogenase for improved

5) Chopra, I., Hawkey, P. M., and Hinton, M. (1992) Tetracyclines,

chlortetracycline biosynthesis in industrial Streptomyces aureofaciens .

Metab. Eng. 19, 69−78.

molecular and clinical aspects. J. Antimicrob. Chemother. 29 , 245 −277.

6) Chopra, I., and Roberts, M. (2001) Tetracycline antibiotics:

mode of action, applications, molecular biology, and epidemiology of

bacterial resistance. Microbiol. Mol. Biol. Rev. 65, 232−60.

7) Brodersen, D. E., Clemons, W. M., Jr., Carter, A. P., Morgan-

Warren, R. J., Wimberly, B. T., and Ramakrishnan, V. (2000) The

structural basis for the action of the antibiotics tetracycline,

pactamycin, and hygromycin B on the 30S ribosomal subunit. Cell

and Jackson, C. J. (2016) Physiology, Biochemistry, and Applications

(24) Greening, C., Ahmed, F. H., Mohamed, A. E., Lee, B. M.,

Pandey, G., Warden, A. C., Scott, C., Oakeshott, J. G., Taylor, M. C.,

(

^o ^f ^F 420 - and F -Dependent Redox Reactions. Microbiol. Mol. Biol. Rev.

(

0, 451−493.

25) Hossain, M. S., Le, C. Q., Joseph, E., Nguyen, T. Q., Johnson-

Winters, K., and Foss, F. W. (2015) Convenient synthesis of

deazaflavin cofactor F and its activity in F 420 -dependent NADP

reductase. Org. Biomol. Chem. 13, 5082−5085.

(

03, 1143−54.

8) Eisen, D. P. Doxycycline. In Kucers’ The Use of Antibiotics: A

Clinical Review of Antibacterial, Antifungal, Antiparasitic, and Antiviral

Drugs, 6th ed.; Lindsay Grayson, M., Cosgrove, S. E., Crowe, S.,

ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology

26) Partow, S., Siewers, V., Bjorn, S., Nielsen, J., and Maury, J.

Articles

(

2010) Characterization of different promoters for designing a new

expression vector in Saccharomyces cerevisiae . Yeast 27, 955−64.

27) Peric-Concha, N., Borovicka, B., Long, P. F., Hranueli, D.,

(45) Woolston, B. M., Edgar, S., and Stephanopoulos, G. (2013)

Metabolic engineering: past and future. Annu. Rev. Chem. Biomol. Eng.

4, 259−88.

(46) Zhou, K., Qiao, K., Edgar, S., and Stephanopoulos, G. (2015)

Distributing a metabolic pathway among a microbial consortium

enhances production of natural products. Nat. Biotechnol. 33 , 377 −83.

Waterman, P. G., and Hunter, I. S. (2005) Ablation of theotcC Gene

Encoding a Post-polyketide Hydroxylase from the Oxytetracyline

Biosynthetic Pathway in Streptomyces rimosus Results in Novel

Polyketides with Altered Chain Length. J. Biol. Chem. 280, 37455−

(47) Wu, X., Zha, J., and Koffas, M. A. G. (2020) Microbial

production of bioactive chemicals for human health. Current Opinion

in Food Science 32, 9−16.

(

7460.

28) McCormick, J. R. D., Miller, P. A., Growich, J. A., Sjolander, N.

(48) Jawed, K., Yazdani, S. S., and Koffas, M. A. G. (2019) Advances

in the development and application of microbial consortia for

metabolic engineering. Metabolic Engineering Communications 9,

e00095.

O., and Doerschuk, A. P. (1958) Two new tetracycline-related

compounds: 7-chloro-5a(11a)-dehydrotetracycline and 5a-epi-tetra-

cycline. a new route to tetracycline. J. Am. Chem. Soc. 80, 5572−5573.

(49) Rogalski, W. Chemical Modiﬁcation of the Tetracyclines. In

29) Eirich, L. D., Vogels, G. D., and Wolfe, R. S. (1979)

The Tetracyclines; Hlavka, J. J., and Boothe, J. H., Eds.; Springer:

Distribution of coenzyme F ₄ ₂ ₀ and properties of its hydrolytic

fragments. J. Bacteriol. 140, 20−7.

(

Berlin, 1985; pp 179−316.

(50) Chooi, Y. H., Cacho, R., and Tang, Y. (2010) Identification of

30) Jacobson, F. S., Daniels, L., Fox, J. A., Walsh, C. T., and Orme-

the viridicatumtoxin and griseofulvin gene clusters from Penicillium

aethiopicum . Chem. Biol. 17, 483−94.

Johnson, W. H. (1982) Purification and properties of an 8-hydroxy-5-

deazaflavin-reducing hydrogenase from Methanobacterium thermoauto-

trophicum . J. Biol. Chem. 257, 3385−8.

(

51) Pirie, C. M., De Mey, M., Prather, K. L. J., and Ajikumar, P. K.

2013) Integrating the Protein and Metabolic Engineering Toolkits

31) Yamazaki, S., Tsai, L., and Stadtman, T. C. (1982) Analogs of

for Next-Generation Chemical Biosynthesis. ACS Chem. Biol. 8, 662−

672.

-hydroxy-5-deazaflavin cofactor: relative activity as substrates for 8-

hydroxy-5-deazaflavin-dependent NADP+ reductase from Methano-

coccus vannielii . Biochemistry 21, 934−939.

(52) Morita, T., and Takegawa, K. (2004) A simple and efficient

procedure for transformation of Schizosaccharomyces pombe. Yeast

21, 613−617.

(

32) Bashiri, G., Squire, C. J., Moreland, N. J., and Baker, E. N.

Hutchison, C. A., 3rd, and Smith, H. O. (2009) Enzymatic assembly

2008) Crystal Structures of F ₄ ₂ ₀ -dependent Glucose-6-phosphate

(53) Gibson, D. G., Young, L., Chuang, R. Y., Venter, J. C.,

of DNA molecules up to several hundred kilobases. Nat. Methods 6,

Dehydrogenase FGD1 Involved in the Activation of the Anti-

tuberculosis Drug Candidate PA-824 Reveal the Basis of Coenzyme

and Substrate Binding. J. Biol. Chem. 283, 17531−17541.

43−5.

54) Green, S. R., and Moehle, C. M. (1999) Media and Culture of

Yeast. Curr. Protoc. Cell Biol. 4, 1.6.1−1.6.12.

33) Le, C. Q., Joseph, E., Nguyen, T., and Johnson-Winters, K.

(

2015) Optimization of Expression and Purification of Recombinant

Archeoglobus fulgidus F ₄ ₂ ₀ H :NADP Oxidoreductase, an F ₄ ₂ ₀

Cofactor Dependent Enzyme. Protein J. 34 , 391 −397.

34) Warkentin, E. (2001) Structures of F H :NADP oxidor-

eductase with and without its substrates bound. EMBO Journal 20,

561−6569.

35) Ohnishi, Y., Ishikawa, J., Hara, H., Suzuki, H., Ikenoya, M.,

(

Ikeda, H., Yamashita, A., Hattori, M., and Horinouchi, S. (2008)

Genome Sequence of the Streptomycin-Producing Microorganism

Streptomyces griseus IFO 13350. J. Bacteriol. 190, 4050−4060.

(

36) Petkovic, H., Cullum, J., Hranueli, D., Hunter, I. S., Peric-

́ ́

Concha, N. a., Pigac, J., Thamchaipenet, A., Vujaklija, D. i., and Long,

P. F. (2006) Genetics of Streptomyces rimosus, the Oxytetracycline

Producer. Microbiol. Mol. Biol. Rev. 70, 704−728.

(

37) Mitscher, L. A. Degradation and Structure Proofs. In The

Chemistry of the Tetracycline Antiobitcs; Marcel Dekker, Inc.: New

York, 1978; pp 122−164.

(

38) Miller, P. A., and Hash, J. H. (1975) Methods Enzymol. 43,

(

06−607.

39) Blackwood, R. K., and English, A. R. Structure−Activity

Relationships in the Tetracycline Series. In Adv. Appl. Microbiol.;

Perlman, D., Ed.; Academic Press, 1970; Vol. 13 , pp 237 −266.

40) Petersen, J. L., and Ronan, P. J. (2010) Critical role of 7,8-

didemethyl-8-hydroxy-5-deazariboflavin for photoreactivation in

Chlamydomonas reinhardtii . J. Biol. Chem. 285, 32467−75.

(

41) Decamps, L., Philmus, B., Benjdia, A., White, R., Begley, T. P.,

and Berteau, O. (2012) Biosynthesis of F , Precursor of the F

420

Cofactor, Requires a Unique Two Radical-SAM Domain Enzyme and

Tyrosine as Substrate. J. Am. Chem. Soc. 134, 18173 −18176.

42) Oltmanns, O., and Bacher, A. (1972) Biosynthesis of

riboflavine in Saccharomyces cerevisiae : the role of genes rib 1 and

rib 7. J. Bacteriol. 110, 818−22.

44) Nogae, I., and Johnston, M. (1990) Isolation and character-

43) Fischer, M., and Bacher, A. (2005) Biosynthesis of

flavocoenzymes. Nat. Prod. Rep. 22, 324.

ization of the ZWF1 gene of Saccharomyces cerevisiae , encoding

glucose-6-phosphate dehydrogenase. Gene 96, 161−169.