**3. Proposed industrial production of C16:1∆6 using a bioreactor**

Microbial production of various materials has a number of advantages over chemical synthesis such as cost effectiveness, ease of production and regiospecific production. Considering the unique function of C16:1Δ6, and its potential therapeutic usefulness, we investigated the possibility of microbiologically producing C16:1Δ6.

There are many reports regarding microbial production of polyunsaturated fatty acids under simple and mild conditions. The production of γ-linoleic acid and arachidonic acid by species of the fungus *Mortierella* is a valuable product for many applications, and the triglyceride esters of poly unsaturated fatty acids that are produced in a fermentation process are used in skin care products (Certik & Shimizu, 1999a; Suzuki, 1987, 1988; Yamada et al., 1987) and baby formula (Certik & Shimizu, 1999b; Shinmen et al., 1989; Yamada, 1988). We isolated an alkane-assimilating *Rhodococcus* sp. strain from a soil sample and found that a mutant was capable of introducing a *cis*-double bond into various aliphatic substrates. We used this mutant for production of C16:1Δ6, investigated the production process, and further constructed a bioreactor system for its industrial production.

Fig. 4. Relationship between free C16:1Δ6 and *S. aureus* in the skin. A: in the skin of AD patients and Healthy Controls; B: in the skin of AD patients who exhibited more than 10 counts/cm2 of *S. aureus*. AD patients; (●), Healthy Controls; (∆)

C16:1Δ6 in AD patients may be one of the factors that contribute to the colonization of *S.* 

We have confirmed that C16:1Δ6 functions as a natural antibacterial component and has very unique properties including "selective antibacterial activity", in which it shows effective antibacterial activity against transient *S. aureus* but not against residential *S. epidermidis*. By topical application of this fatty acid to AD patients, the colonization of *S. aureus,* which is believed to be an exacerbation factor for this disease, was repressed, thereby indicating the effectiveness of this fatty acid. From the results of these investigations, a new approach for improvement of skin disorders, based on normalization of the microflora on human skin, was proposed using the selective

Microbial production of various materials has a number of advantages over chemical synthesis such as cost effectiveness, ease of production and regiospecific production. Considering the unique function of C16:1Δ6, and its potential therapeutic usefulness, we

There are many reports regarding microbial production of polyunsaturated fatty acids under simple and mild conditions. The production of γ-linoleic acid and arachidonic acid by species of the fungus *Mortierella* is a valuable product for many applications, and the triglyceride esters of poly unsaturated fatty acids that are produced in a fermentation process are used in skin care products (Certik & Shimizu, 1999a; Suzuki, 1987, 1988; Yamada et al., 1987) and baby formula (Certik & Shimizu, 1999b; Shinmen et al., 1989; Yamada, 1988). We isolated an alkane-assimilating *Rhodococcus* sp. strain from a soil sample and found that a mutant was capable of introducing a *cis*-double bond into various aliphatic substrates. We used this mutant for production of C16:1Δ6, investigated the production process, and further constructed a bioreactor system for its industrial

**B**

**Number of** *S.aureus* **(counts/cm2**

Fig. 4. Relationship between free C16:1Δ6 and *S. aureus* in the skin. A: in the skin of AD patients and Healthy Controls; B: in the skin of AD patients who exhibited more than 10

**)**

**104**

**103**

**102**

**101**

**1**

Log(Number of *S.aureus*)=-2.28497[Free C16:1Δ6]+3.42416 R2=0.40998

**Free C16:1**Δ**6(**μ**g/cm2**

**0 0.2 0.4 0.6 0.8 1.0**

**)**

**3. Proposed industrial production of C16:1∆6 using a bioreactor** 

investigated the possibility of microbiologically producing C16:1Δ6.

**Healthy controls**

counts/cm2 of *S. aureus*. AD patients; (●), Healthy Controls; (∆)

**Free C16:1**Δ**6(**μ**g/cm2) 0 2 4 6**

*aureus* (Takigawa et al., 2005).

antibacterial activity of C16:1Δ6.

production.

**3000**

**2000**

**1000**

**Number of** *S.aureus* **(counts/cm2**

**)**

**0**

**A**

**AD patients**

#### **3.1 Desaturation reaction of aliphatic substrates by a mutant strain of alkaneassimilating** *Rhodococcus* **sp.**

We found that the double mutant designated *Rhodococcus* sp. KSM-MT66 had the ability to desaturate various aliphatic compounds such as alkanes, chloro alkanes and fatty acid esters, and that this mutant strain produced unsaturated compounds extracellularly. When hexadecanoic acid esters such as methyl, propyl, isopropyl, and isobutylester were supplied as substrates to the resting cells, their corresponding *cis*-desaturated compounds were produced at a concentration of 0.5, 20, 53 and 7 g/l in 3 days. The enzyme(s) responsible for the desaturation reaction appears to recognize mainly the sixth carbon from the carbonyl carbons (Fig. 5). The bioproduction of C16:1Δ6 esters has not been previously reported (Koike et al., 1999). When alkanes were supplied to this mutant strain, the main products had a double bond at the ninth carbon from the terminal methyl group. These experiments indicated that the unsaturated position differed according to differences in the supplied substrates (Koike et al., 2000a).

Fig. 5. The patterns of regiospecific desaturation of aliphatic substrates by *Rhodococcus* sp. strain KSM-MT66 cells.

#### **3.2 Production of a C16:1∆6 ester by resting cells of** *Rhodococcus* **sp. KSM-MT66**

We determined the condition with resting cells of the *Rhodococcus* sp. KSM-MT66 strain to produce C16:1Δ6. The reaction mixture contained 20% (w/v) hexadecanoic acid isopropyl ester (IP-C16:0), 0.25 M phosphate buffer (pH 7.0), 1.0% (w/v) monosodium glutamate, 2

Improvement of Atopic Dermatitis by Human Sebaceous Fatty Acids and Related Lipids 315

Fig. 7. Repeat-batch reaction using a membrane bioreactor system with a phase inversion for

The resting cell production of IP-C16:1Δ6 described above includes four complicated steps; (i) cell growth, (ii) cell harvesting, (iii) incubation of cells with substrate and (iv) phase-

In order to develop a more convenient production process we therefore investigated production of IP-C16:1Δ6 using a fermentation process. We first established a basal medium for *Rhodococcus* sp. KSM-MT66. When production was performed under resting cell conditions, 18 g/l of IP-C16:1Δ6 was produced over 3 days of cultivation. Optimization of the concentrations of metal ions greatly improved production of IP-C16:1Δ6, resulting in

**3.5 Fermentative production of IP-C16:1∆6 using the mutant** *Rhodococcus* **sp. KSM-**

Since there was a possibility that the IP-C16:1Δ6 product might be degraded by esterases present in the *Rhodococcus* sp. KSM-MT66, we attempted to create *Rhodococcus* mutants with

the production of IP-C16:1Δ6 by the resting cells of *Rhodococcus* sp. KSM-MT66.

*Rhodococcus* **sp. KSM-MT66** 

**T64** 

**3.4 Fermentative production of IP-C16:1∆6 using growing cells of the mutant** 

inversed separation of the product using a hydrophobic fiber membrane system.

production of 52 g/l over 4 days of cultivation (see sec. 3.5).

mM thiamine, 2 mM MgSO4, and 5% (wet w/v) resting cells. Bioconversion was performed using a 1.0 l working volume in a 2.6 l bioreactor at 26 ºC with aeration and agitation. Glutamate, thiamine, and MgSO4 prevented cell damage of this strain during repeat-batch bioconversion, and optimum concentrations of these factors were maintained in the reaction mixture. Under optimum conditions, about 50 g/l of C16:1Δ6 isopropyl ester (IP-C16:1Δ6) was produced extracellularly in 3 days.

Since this reaction mixture consisted of water, oil and cells, we considered that this may make it difficult to separate the products. We therefore designed a new process to recover the products. Thus, the oil in water (O/W-type) emulsified reaction mixture was kept without agitation for 20 hours, the water layer was drained out, an appropriate volume of substrate, IP-C16:0, was added to the reaction mixture to invert the emulsion phase to a W/O-type, and the product, IP-C16:1Δ6, in the continuous oil phase was recovered through a hydrophobic hollow-fiber module (Fig. 6). The hydrophobic cells were concentrated in the oil phase. It was then possible to start the next batch by adding fresh medium to these cells (Koike et al., 2000b; Takeuchi et al., 1990). This system allows repeat-batch reactions without having to use an organic solvent to recover the products.

Fig. 6. Schematic diagram of the membrane bioreactor system for Rhodococcal bioconversion. (A) Reactor vessel; (B) Hydrophobic hollow-fiber module; (C) Circulation pump.

#### **3.3 Repeat-batch production of the C16:1∆6 ester**

The reason for the decline in late production during repeat batch production may be due to various factors such as accumulation of some inhibitory products, cell damage and/or exhaustion of some nutrients. Since monosodium glutamate was found to reduce cell damage, we investigated the effect of monitoring and adjusting the concentration of monosodium glutamate on productivity during repeat batch production (see sec. 3-2). The results of this experiment showed that a productivity of 0.8 g/l was attained over the course of 13 cycles (300 hours) by maintaining the concentration of monosodium glutamate between 0.5 and 1.5% (Fig. 7). It was possible to recover the produced IP-C16:1Δ6 using the hollow-fiber system, and IP-C16:1Δ6 was then purified by urea adduct treatment, evaporated and applied to a silica gel column. The total yield of IP-C16:1Δ6 using this procedure was 79% and the purity was over 97%. Long term operation of C16:1Δ6 ester production was achieved using a 2.6 l fermentor (Koike et al., 2000b).

mM thiamine, 2 mM MgSO4, and 5% (wet w/v) resting cells. Bioconversion was performed using a 1.0 l working volume in a 2.6 l bioreactor at 26 ºC with aeration and agitation. Glutamate, thiamine, and MgSO4 prevented cell damage of this strain during repeat-batch bioconversion, and optimum concentrations of these factors were maintained in the reaction mixture. Under optimum conditions, about 50 g/l of C16:1Δ6 isopropyl ester (IP-C16:1Δ6)

Since this reaction mixture consisted of water, oil and cells, we considered that this may make it difficult to separate the products. We therefore designed a new process to recover the products. Thus, the oil in water (O/W-type) emulsified reaction mixture was kept without agitation for 20 hours, the water layer was drained out, an appropriate volume of substrate, IP-C16:0, was added to the reaction mixture to invert the emulsion phase to a W/O-type, and the product, IP-C16:1Δ6, in the continuous oil phase was recovered through a hydrophobic hollow-fiber module (Fig. 6). The hydrophobic cells were concentrated in the oil phase. It was then possible to start the next batch by adding fresh medium to these cells (Koike et al., 2000b; Takeuchi et al., 1990). This system allows repeat-batch reactions without

was produced extracellularly in 3 days.

pump.

having to use an organic solvent to recover the products.

**3.3 Repeat-batch production of the C16:1∆6 ester** 

Fig. 6. Schematic diagram of the membrane bioreactor system for Rhodococcal

production was achieved using a 2.6 l fermentor (Koike et al., 2000b).

bioconversion. (A) Reactor vessel; (B) Hydrophobic hollow-fiber module; (C) Circulation

The reason for the decline in late production during repeat batch production may be due to various factors such as accumulation of some inhibitory products, cell damage and/or exhaustion of some nutrients. Since monosodium glutamate was found to reduce cell damage, we investigated the effect of monitoring and adjusting the concentration of monosodium glutamate on productivity during repeat batch production (see sec. 3-2). The results of this experiment showed that a productivity of 0.8 g/l was attained over the course of 13 cycles (300 hours) by maintaining the concentration of monosodium glutamate between 0.5 and 1.5% (Fig. 7). It was possible to recover the produced IP-C16:1Δ6 using the hollow-fiber system, and IP-C16:1Δ6 was then purified by urea adduct treatment, evaporated and applied to a silica gel column. The total yield of IP-C16:1Δ6 using this procedure was 79% and the purity was over 97%. Long term operation of C16:1Δ6 ester

Fig. 7. Repeat-batch reaction using a membrane bioreactor system with a phase inversion for the production of IP-C16:1Δ6 by the resting cells of *Rhodococcus* sp. KSM-MT66.

#### **3.4 Fermentative production of IP-C16:1∆6 using growing cells of the mutant**  *Rhodococcus* **sp. KSM-MT66**

The resting cell production of IP-C16:1Δ6 described above includes four complicated steps; (i) cell growth, (ii) cell harvesting, (iii) incubation of cells with substrate and (iv) phaseinversed separation of the product using a hydrophobic fiber membrane system.

In order to develop a more convenient production process we therefore investigated production of IP-C16:1Δ6 using a fermentation process. We first established a basal medium for *Rhodococcus* sp. KSM-MT66. When production was performed under resting cell conditions, 18 g/l of IP-C16:1Δ6 was produced over 3 days of cultivation. Optimization of the concentrations of metal ions greatly improved production of IP-C16:1Δ6, resulting in production of 52 g/l over 4 days of cultivation (see sec. 3.5).

#### **3.5 Fermentative production of IP-C16:1∆6 using the mutant** *Rhodococcus* **sp. KSM-T64**

Since there was a possibility that the IP-C16:1Δ6 product might be degraded by esterases present in the *Rhodococcus* sp. KSM-MT66, we attempted to create *Rhodococcus* mutants with

Improvement of Atopic Dermatitis by Human Sebaceous Fatty Acids and Related Lipids 317

Delta 6-desaturases have been reported to be obtained from animals, plants, fungi and cyanobacteria (Aki et al., 1999; Cahoon et al., 1994; Inagaki et al., 2003; Okayasu et al., 1981; Reddy & Thomas, 1996; Zhan et al., 2004). Human and rat fatty acid desaturase 2 (FADS2) -encoding delta 6-desaturases recognize a saturated fatty acid as a substrate under some conditions (Ge et al., 2003; Guillou et al., 2003), and a human delta 6 desaturase gene has been reported to be expressed only in the sebaceous gland. Regarding microbial delta 6-desaturases, industrial producers of unsaturated fatty acids have focused mainly on fungi of the *Mortierella* spp., especially *Mortierella alpina* and *Mortierella cincinelloides*. Two types of delta 6-desaturases have been reported to be expressed by *Mortierella alpina.* The corresponding genes were isolated and the sequences and functions of these genes have been analyzed (Huang et al., 1999; Sakuradani et al., 1999; Sakuradani & Shimizu, 2003). However, even though many delta 6-desaturase genes have been

obtained, no delta 6-desaturase genes have been purified from *Rhodococcus* sp.

which shows 40% of the esterase activity of the parent strain, KSM-MT66.

**IP-C16:1**

KSM-T64 was used.

Δ**6 , MSG (g/L)**

We aimed to obtain delta 6-desaturase genes from the *Rhodococcus* sp. KSM-T64 strain,

**0 20 40 60 80 100**

**Cultivation time (h)**

IP-C16:1Δ6 (●), monosodium glutamate (MSG) (○). Cultivation was performed at 26 ºC, with agitation at 350 rpm, aeration at 0.3 vvm and pressure at 0.2 kg/cm2. The *Rhodococcus* sp.

**4.2 Cloning of delta 6-desaturase genes from the** *Rhodococcus* **sp. KSM-T64 strain**  The production of IP-C16:1Δ6 by *Rhodococcus* sp. was improved by the addition of metal ions as mentioned above. We therefore targeted the membrane-bound delta 6-desaturase and attempted to clone a gene whose encoded protein could introduce a *cis*-double bond**.**  We cloned the delta 6-desaturase gene on the basis of previously reported conserved sequences. Histidine motifs (Shanklin & Fox, 1994; Shanklin, 2009) that bind to ferric ions and whose sequences are known to be conserved in membrane-bound desaturases in a

Fig. 8. Production of IP-C16:1Δ6 in a 30 l-jar fermentor under optimized conditions.

reduced esterase activity by UV irradiation. Of the colonies which showed lower growth than KSM-MT66 on the minimum agar containing IP-C16:0, one mutant, designated KSM-T64, displayed 40% of the esterase activity of KSM-MT66.

Using this mutant strain T64, and optimizing culture conditions, more than 60 g/l of IP-C16:1Δ6 could be produced in a flask (Table 1). Optimization of culture conditions in a 30 l jar fermentor, resulted in production of 50 g/l over 4 days of cultivation (Fig. 8). Furthermore, C16:1Δ6 can be easily obtained by simple hydrolysis of IP-C16:1Δ6 (Araki et al., 2007).


a Not added.

Table 1. Production of IP-C16:1Δ6 by *Rhodococcus* sp. KSM-MT66 and KSM-T64.
