Genetic Determinants of Hydroxycinnamic Acid Metabolism in ...the Lactobacillus species observed in this study, genotype and phenotype are com-pared in Table 2. In lactobacilli, genotype

Genetic Determinants of Hydroxycinnamic Acid Metabolism inHeterofermentative Lactobacilli

Gautam Gaur,a Jee-Hwan Oh,b Pasquale Filannino,c Marco Gobbetti,d Jan-Peter van Pijkeren,b Michael G. Gänzlea,e

aUniversity of Alberta, Department of Agricultural, Food and Nutritional Science, Edmonton, Alberta, CanadabDepartment of Food Science, University of Wisconsin—Madison, Madison, Wisconsin, USAcDepartment of Soil, Plant and Food Science, University of Bari Aldo Moro, Bari, ItalydFaculty of Sciences and Technology, Free University of Bozen-Bolzano, Bolzano, ItalyeHubei University of Technology, College of Bioengineering and Food Science, Wuhan, Hubei, People’s Republic of China

ABSTRACT Phenolic acids are among the most abundant phenolic compounds inedible parts of plants. Lactic acid bacteria (LAB) metabolize phenolic acids, but theenzyme responsible for reducing hydroxycinnamic acids to phenylpropionic acids(HcrB) was only recently characterized in Lactobacillus plantarum. In this study, het-erofermentative LAB species were screened for their hydroxycinnamic acid metabo-lism. Data on strain-specific metabolism in combination with comparative genomicanalyses identified homologs of HcrB as putative phenolic acid reductases. Par1 andHcrF both encode putative multidomain proteins with 25% and 63% amino acididentity to HcrB, respectively. Of these genes, par1 in L. rossiae and hcrF in L. fermen-tum were overexpressed in response to hydroxycinnamic acids. The deletion of par1in L. rossiae led to the loss of phenolic acid metabolism. The strain-specific metabo-lism of phenolic acids was congruent with the genotype of lactobacilli; however,phenolic acid reductases were not identified in strains of Weissella cibaria that re-duced hydroxycinnamic acids to phenylpropionic acids. Phylogenetic analysis of ma-jor genes involved in hydroxycinnamic acid metabolism in strains of the genus Lac-tobacillus revealed that Par1 was found to be the most widely distributed phenolicacid reductase, while HcrB was the least abundant, present in less than 9% of Lacto-bacillus spp. In conclusion, this study increased the knowledge on the genetic deter-minants of hydroxycinnamic acid metabolism, explaining the species- and strain-specific metabolic variations in lactobacilli and providing evidence of additionalenzymes involved in hydroxycinnamic acid metabolism of lactobacilli.

IMPORTANCE The metabolism of secondary plant metabolites, including phenoliccompounds, by food-fermenting lactobacilli is a significant contributor to the safety,quality, and nutritional quality of fermented foods. The enzymes mediating hydroly-sis, reduction, and decarboxylation of phenolic acid esters and phenolic acids in lac-tobacilli, however, are not fully characterized. The genomic analyses presented hereprovide evidence for three novel putative phenolic acid reductases. Matching com-parative genomic analyses with phenotypic analysis and quantification of gene ex-pression indicates that two of the three putative phenolic acid reductases, Par1 andHcrF, are involved in reduction of hydroxycinnamic acids to phenylpropionic acids;however, the activity of Par2 may be unrelated to phenolic acids and recognizesother secondary plant metabolites. These findings expand our knowledge on themetabolic potential of lactobacilli and facilitate future studies on activity and sub-strate specificity of enzymes involved in metabolism of phenolic compounds.

KEYWORDS phenolic acid reductase, phenolic acid decarboxylase,heterofermentative lactobacilli, metabolism, Lactobacillus, Lactobacillus fermentum,phenolic compounds

Citation Gaur G, Oh J-H, Filannino P, GobbettiM, van Pijkeren J-P, Gänzle MG. 2020. Geneticdeterminants of hydroxycinnamic acidmetabolism in heterofermentative lactobacilli.Appl Environ Microbiol 86:e02461-19. https://doi.org/10.1128/AEM.02461-19.

Editor Donald W. Schaffner, Rutgers, The StateUniversity of New Jersey

Copyright © 2020 American Society forMicrobiology. All Rights Reserved.

Address correspondence to Michael G. Gänzle,[email protected].

Received 23 October 2019Accepted 16 December 2019

Accepted manuscript posted online 20December 2019Published

FOOD MICROBIOLOGY

crossm

March 2020 Volume 86 Issue 5 e02461-19 aem.asm.org 1Applied and Environmental Microbiology

18 February 2020

on March 30, 2021 by guest

http://aem.asm

.org/D

ownloaded from

https://orcid.org/0000-0002-1235-5138https://orcid.org/0000-0003-0972-928Xhttps://doi.org/10.1128/AEM.02461-19https://doi.org/10.1128/AEM.02461-19https://doi.org/10.1128/ASMCopyrightv2mailto:[email protected]://crossmark.crossref.org/dialog/?doi=10.1128/AEM.02461-19&domain=pdf&date_stamp=2019-12-20https://aem.asm.orghttp://aem.asm.org/

Phenolic acids are a class of phenolic compounds that are abundant in edible partsof plants. In plants, hydroxycinnamic acids occur mainly bound to cell wall poly-saccharides, as glycosides or as esters (1–4). Phenolic acids can be distinguished as twotypes: hydroxybenzoic acids and the more abundant hydroxycinnamic acids. Epidemi-ological studies have associated consumption of dietary fibers rich in phenolic acidswith reduction in chronic inflammation and a reduced risk of colon cancer, type 2diabetes, neurodegenerative diseases, and cardiovascular diseases (5, 6). The freehydroxycinnamic acids caffeic and ferulic acids display anti-inflammatory and antican-cer properties with reduction in tumor multiplicity (7, 8). Hydroxycinnamic acids alsohave antimicrobial activity at concentrations that correlate to their abundance in plants(9).

During fermentation of plants, esterase activities of lactic acid bacteria releasebound hydroxycinnamic acids. In addition, lactic metabolism converts hydroxycinnamicacids via decarboxylation and reduction reactions (3, 10). The decarboxylation as wellas the reduction of hydroxycinnamic acids by microbial metabolism reduces theirantimicrobial activity (9). Vinyl as well as ethyl derivatives that result from microbialmetabolism are flavor volatiles that impact the aroma of fermented foods (6). Vinylderivatives also react with anthocyanins and 3-deoxyanthocyanins that are presentin many fruits and sorghum, respectively, to form pyranoanthocyanins and 3-deoxypyranoanthocyanins (11, 12). The conversion of phenolic acids during foodfermentations thus impacts quality, safety, and nutritional properties of fermentedfoods. Metabolism of phenolic acids, however, is strain specific and relatively poorlycharacterized compared to other metabolic pathways of lactic acid bacteria.

Enzymes responsible for metabolism of hydroxycinnamic acids have been identifiedprimarily in homofermentative lactic acid bacteria. Phenolic acid esterases were char-acterized in Lactobacillus johnsonii (13) and Lactobacillus plantarum (14, 15). Phenolicacid decarboxylases were characterized only in L. plantarum (16, 17). Until recently,Clostridium was the closest relative to Lactobacillaceae for which enzymes capable ofreducing hydroxycinnamates were characterized (18, 19). In 2018, phenolic acid reduc-tases reducing hydroxycinnamic acids to substituted phenylpropionic acids (HcrB) orreducing the decarboxylated vinyl derivatives to ethyl derivatives (VrpA) were charac-terized in L. plantarum (20, 21). Heterofermentative lactic acid bacteria, including L.fermentum (9, 22), L. rossiae (23, 24), L. kunkeei (25), and Weissella cibaria (23) also reducehydroxycinnamic acids to the corresponding phenylpropionic acids; however, enzymesinvolved in phenolic acid metabolism of heterofermentative lactic acid bacteria havenot been characterized (26). In heterofermentative lactobacilli, the reduction of phe-nolic acids reduces NADH; this conversion increases the energy yield in the phospho-ketolase pathway and thus has a high priority in heterolactic metabolism (23, 27).Because heterofermentative lactobacilli differ phylogenetically and physiologically fromhomofermentative lactobacilli (28), these organisms may harbor novel genes andenzymes responsible for reduction of hydroxycinnamic acids.

This work aimed to study hydroxycinnamic metabolism of eight heterofermentativestrains of lactic acid bacteria, using two L. plantarum strains as a reference. Thestrain-specific phenolic acid metabolism guided a comparative genomic analyses toidentify homologs to HcrB as putative phenolic acid reductases. Two novel phenolicacid reductases were identified in heterofermentative lactobacilli. Further bioinformat-ics analyses also determined the presence of different phenolic acid reductases, phe-nolic acid decarboxylase and vinyl phenol reductase, across Lactobacillaceae.

RESULTSHydroxycinnamic acid metabolism of heterofermentative lactobacilli. To assess

the metabolism of hydroxycinnamic acids by heterofermentative lactobacilli, strainswere grown in the presence of sinapic acid, ferulic acid, or caffeic acid and metaboliteswere analyzed by reverse-phase high-performance liquid chromatography (HPLC) (Fig.1). Sinapic acid was metabolized exclusively by reduction, while ferulic and caffeic acidswere converted by strain- or species-specific reduction and/or decarboxylation reac-

Gaur et al. Applied and Environmental Microbiology

March 2020 Volume 86 Issue 5 e02461-19 aem.asm.org 2


http://aem.asm

.org/D

ownloaded from

https://aem.asm.orghttp://aem.asm.org/

tions (Fig. 1). L. plantarum TMW1.460 and L. plantarum FUA3584 reduced sinapic acidto dihydrosinapic acid but differed with respect to the metabolism of ferulic and caffeicacids. L. plantarum TMW1.460 metabolized ferulic acid to dihydroferulic acid andvinylguaiacol by reduction and decarboxylation reactions, respectively, whereas thedecarboxylation product vinylcatechol was the only metabolite from caffeic acid. L.plantarum FUA3584 reduced ferulic acid but decarboxylated caffeic acid. Vinylcatecholwas not detected and its reduced form, ethylcatechol, was the only product (data notshown). L. rossiae strains also differed in metabolism of caffeic acid. L. rossiae C5reduced all three substrates, while L. rossiae FUA3583 reduced sinapic and ferulic acidsbut metabolized caffeic acid mainly by decarboxylation. Metabolites observed incultures of L. fermentum FUA3589 and W. cibaria 10M were similar to those for L. rossiaeC5, showing exclusive reduction of all three substrates. L. hammesii and L. brevis strainssolely displayed decarboxylation activity toward ferulic and caffeic acids and did notmetabolize sinapic acid. L. reuteri DSM20016 and L. kunkeei DSM12361 showed noactivity against any of the three substrates.

Identification of putative phenolic acid reductases. Putative phenolic acid re-ductases in L. rossiae and L. fermentum strains were identified by searching for ho-mologs of phenolic acid reductase HcrB in L. plantarum WCFS1 (21); other proteins thatwere encoded in the proposed operon, including HcrR, HcrA, and HcrC, were addition-ally used as query sequences (Table 1). L. plantarum TMW1.460 and FUA3584 containedthe complete operon including the genes encoding HcrR, HcrA, HcrB, and HcrC; aminoacid similarities to the protein sequences in L. plantarum WCFS1 were 98% or higher.L. fermentum also harbored an HcrR homolog (Table 1). In L. rossiae C5 and FUA3583,a LysR-type transcriptional regulator encoded by parR with 28% amino acid identity toHcrR was present (Table 1). HcrR homologs in L. rossiae and L. fermentum were encodedin proximity of putative phenolic acid reductases. Homologs of HcrA were present in allreductase-positive strains as well in L. reuteri DSM 20016 and W. cibaria 10M. Two HrcBhomologs, Par1 and Par2, were identified in L. rossiae C5 and L. rossiae FUA3583, butthese proteins exhibited an amino acid identity of only 25 to 26% to HcrB. Of note, Par1and Par2 homologs were also present in both L. plantarum strains. An HcrB homolog inL. fermentum FUA3589, termed HcrF, is 63% identical to HcrB but has the same size asPar1/Par2 (Table 1). A protein annotated as flavocytochrome c containing the fumarate

Con

cent

ratio

n (m

M) /

Rel

ativ

e Pe

ak A

rea

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4Dihydrosinapic acid Dihydroferulic acid Vinylguaiacol Dihydrocaffeic acid Vinylcatechol

L. pla

ntaru

m TM

W 1.

460

L. pla

ntaru

m FU

A 35

84

L. ro

ssiae

FUA

3583

L. ro

ssiae

C5

L. fer

mentu

m FU

A 35

83

L. ha

mmes

ii DSM

1638

1

L. br

evis

TMW

1.46

5

L. re

uteri D

SM 20

016

L. ku

nkee

i DSM

1236

1

Weis

sella

ciba

ria 10

MFIG 1 Metabolites of phenolic acid conversion by strains of Lactobacillus and Weissella after incubationwith 1 mM concentrations of different hydroxycinnamic acids. Data are shown as means � SDs from twoindependent experiments.

Hydroxycinnamic Acid Metabolism in Lactobacilli Applied and Environmental Microbiology



http://aem.asm

.org/D

ownloaded from


reductase flavoprotein subunit FccA (29) was present in all strains except L. hammesiiand L. brevis. The amino acid identity of all of the FccA proteins to HcrB was less than34%, and the protein length was only 462 to 464 amino acids. HcrC was present onlyin L. plantarum (Table 1).

A potential role of HcrB homologs in phenolic acid metabolism was further assessedby comparison of the domain architecture of the proteins. HcrB consists of 3 domains,an NADH binding domain, a flavin mononucleotide (FMN) binding domain, and a flavinadenine dinucleotide (FAD) binding domain (Fig. 2). Par1/Par2 and HcrF contained twoof the HcrB domains; the domain organization in HcrF matches that of HcrB, but thedomain organization in Par1 and Par2 differs from HcrB. The fumarate reductaseflavoprotein subunit FccA contained only one FAD binding protein domain. HcrR andParR proteins belong to the same LysR transcriptional regulator family.

Expression profile of putative phenolic acid reductases. To further elucidate apotential role of putative phenolic acid reductases role in hydroxycinnamic metabolism,the gene expression by exponentially growing cells in response to phenolic acids wasquantified by reverse transcription-quantitative PCR (RT-qPCR). L. plantarum TMW1.460and FUA3584 overexpressed (P � 0.05) hcrA, hcrB, and hcrC in the presence of hydroxy-cinnamic acids at early exponential phase (Fig. 3). The genes coding for phenolic aciddecarboxylase activity were overexpressed in response to ferulic and caffeic acids; othergenes, including the homologs of par1 and par2 (both strains) and vrpA in L. plantarumFUA3584, were not overexpressed (Fig. 3).

TABLE 1 In silico identification of putative phenolic acid reductases using L. plantarum WCFS1 protein sequences as queries

Bacterial strain

L. plantarum WCFS1 query protein

HcrR,a 315 aa HcrA,a 204 aa HcrB,a 812 aa

Proteinname

Amino acididentity (%)

Proteinlength (aa) Protein name


Proteinlength (aa)

Proteinname


Proteinlength (aa)

L. rossiae C5b ParR 28 304 HcrA2/HcrA3 52/47 202/452 Par1/Par2 25/26 614/616FccA 33 465

L. fermentum FUA3589 HcrR 43 308 HcrA2/HcrA3 52/49 203/450 HcrF 63 617FccA 33 464

L. hammesii DSM16381L. brevis TMW1.465L. reuteri DSM20016 HcrA2 39 190 FccA 34 464

HcrA3 46 416L. kunkeei DSM12361 FccA 30 462W. cibaria 10M HcrA2 50 283 FccA 32 464aProtein accession numbers for HcrR, HrcA, and HcrB are YP_004889274.1, YP_004889275.1, and YP_004889276.1, respectively. aa, amino acids.bSimilar results were obtained for L. rossiae C5 and FUA3583; only results for L. rossiae C5 are shown.

FIG 2 Domain architecture of protein sequences of putative phenolic acid reductases with respect to L.plantarum WCFS1 HcrB. InterPro domain names along with their InterPro identifiers are provided. Thesame colors represent conserved domains across sequences.




http://aem.asm

.org/D

ownloaded from

https://www.ncbi.nlm.nih.gov/protein/YP_004889274.1https://www.ncbi.nlm.nih.gov/protein/YP_004889275.1https://www.ncbi.nlm.nih.gov/protein/YP_004889276.1https://aem.asm.orghttp://aem.asm.org/

L. rossiae C5 and FUA3583 both overexpressed par1 in the presence of all threesubstrates (P � 0.05) (Fig. 4). Interestingly, its homolog par2 was overexpressed 20-fold(P � 0.05) in L. rossiae FUA3583 when cultured with sinapic acid but not in L. rossiae C5.The padR and pad genes of L. rossiae FUA3583 were overexpressed in the presence offerulic and caffeic acids.

L. fermentum FUA3589 overexpressed (P � 0.05) hcrF in the presence of caffeic acid(Fig. 5). Surprisingly, pad was not overexpressed in the presence of any of the threesubstrates. The genes hcrA2, hcrA3, and fccA were never overexpressed under anycondition, and differentially expressed genes were not found in W. cibaria 10M or inreductase-negative strains when cultured with hydroxycinnamic acids.

Phylogenetic analysis of major genes involved in hydroxycinnamic acid me-tabolism. We analyzed the presence of 3 types of phenolic acid reductases, thosecorresponding to hcrB, hcrF, and par1/par2, along with the only known phenolic aciddecarboxylase (corresponding to pad) and the recently characterized vinylphenolreductase (corresponding to vprA), across all the sequenced Lactobacillus spp. andPediococcus spp. Out of the 196 species analyzed, 98 Lactobacillus spp. contained atleast one phenolic acid reductase (Fig. 6A). The most abundant phenolic acid reductasewas that corresponding to par1/par2, which is present in 68 Lactobacillus spp., whilethat corresponding to hcrB was the least abundant and present in only 16 Lactobacillusspp. Phenolic acid reductases were not identified in any of the Pediococcus spp.

FIG 3 Relative fold gene expressions (log2 transformed) of L. plantarum TMW1.460 (top) and L. plantarumFUA3584 (bottom) genes. Strains were incubated with 1 mM concentrations of different hydroxycin-namic acids in mMRS broth, and broth without any hydroxycinnamic acids was used as the referencecondition. An asterisk indicates that the gene is significantly overexpressed (P � 0.05) with respect to itsexpression under the reference condition. Genes and their corresponding proteins or other descriptionsare as follows: hcrR, regulator of hcr operon; hcrABC, phenolic acid reductase; parR, putative regulatorsof phenolic acid reductase; par1 and par2, putative phenolic acid reductase; hcrA2, homolog of hcrA thatis not part of a phenolic acid reductase operon; fccA, subunit of fumarate reductase; padR, phenolic aciddecarboxylase transcriptional regulator; pad, phenolic acid decarboxylase, and vprA, vinyl phenolreductase.




http://aem.asm

.org/D

ownloaded from


Thirty-four Lactobacillus spp. contained more than one type of phenolic acid reductase.Two species, L. plantarum and L. pentosus, contained homologs for all three types ofphenolic acid reductases. L. hayakitensis was the only species which had both hcrB andhcrF phenolic acid reductases, while 5 species encoded hcrB and par1/par2 phenolicacid reductases. The most abundant combination in species having more than one typeof phenolic acid reductase was hcrF and par1/par2, with 26 Lactobacillus spp. havingstrains with both the genes.

Sixty-nine Lactobacillus spp. and 7 Pediococcus spp. carried pad (Fig. 6B). Thecorresponding protein sequence had a very high degree of conservation; proteinsidentified in lactobacilli were more than 76% identical to Pad in L. plantarum WCFS1.The only exception is Pad from L. florum, which had 60% amino acid identity with thequery sequence (data not shown). VrpA was present in 34 Lactobacillus spp. (Fig. 6C).Interestingly, 15 Lactobacillus spp. encoded VrpA but not Pad. Only 19 Lactobacillus spp.possess genes responsible both for decarboxylation of hydroxycinnamic acids to vinylderivatives and for further reducing them to their ethyl derivatives.

Comparison of genotype and phenotype in lactobacilli. To determine whetherthe three putative novel phenolic acid reductases explain phenolic acid metabolism inthe Lactobacillus species observed in this study, genotype and phenotype are com-pared in Table 2. In lactobacilli, genotype and phenotype always matched, i.e., themetabolism of a certain compound was accurately predicted by the presence or

FIG 4 Relative fold gene expressions (log2 transformed) of L. rossiae C5 (top) and L. rossiae FUA3583(bottom) genes. Strains were incubated with 1 mM concentrations of different hydroxycinnamic acids inmMRS broth, and broth without any hydroxycinnamic acids was used as the reference condition. Anasterisk indicates that the gene is significantly overexpressed (P � 0.05) with respect to its expressionunder the reference condition. hcrA3 is a homolog of hcrA that is not part of a phenolic acid reductaseoperon; for descriptions of the rest of the genes, see the legend to Fig. 3.




http://aem.asm

.org/D

ownloaded from


absence of the metabolic genes (Table 2). Exceptions pertained only to the alternativemetabolism by decarboxylation or reduction of hydroxycinnamic acids. L. rossiaeFUA3583 and L. fermentum FUA3589 possessed decarboxylase and reductase geneswith activity on caffeic and ferulic acids; for ferulic acid, however, only the reduceddihydrophenolic acid was observed. This may indicate that the decarboxylase is notactive or not expressed, or it may reflect the strong preference of heterofermentativelactobacilli for cofactor regeneration (23). L. fermentum FUA3589 exclusively reduced allsubstrates, but pad was not overexpressed in the presence of hydroxycinnamic acids.Of note, the vprA-positive strain L. plantarum FUA3584 reduced vinylcatechol but not

Gene Name

hcrR

hcrA

2

hcrA

3

hcrF

fccA

padR pa

d

hcrA

2-R

hcrA

3-R

fccA

-R

fccA

-K

Rel

ative

Fol

d G

ene

Expr

essi

on(L

og2

Tran

sfor

med

)

-4

-3

-2

-1

0

1

2

3

4Sinapic Acid Ferulic Acid Caffeic Acid

*L. fermentum FUA3589 L. reuteri DSM 20016 L. kunkeei

DSM 12361

FIG 5 Relative fold gene expressions (log2 transformed) of L. fermentum FUA3589, L. reuteri DSM20016,and L. kunkeei DSM12361 genes. Strains were incubated with 1 mM concentrations of different hydroxy-cinnamic acids in mMRS broth, and broth without any hydroxycinnamic acids was used as the referencecondition. An asterisk indicates that the gene is significantly overexpressed (P � 0.05) with respect to isexpression under the reference condition. hcrF encodes a putative phenolic acid reductase (HcrBhomolog) in L. fermentum; for descriptions of the rest of the genes, see the legend to Fig. 3.

FIG 6 Phylogenetic analysis of different genes involved in phenolic acid metabolism in 196 Lactobacillus and Pediococcus spp. The name of the species followedby the NCBI protein accession number is provided. Homofermentative and heterofermentative species are represented by red and blue colors, respectively. Thecolor strip represents the lifestyle for species from the work of Duar et al. (34). (A) Phylogenetic analysis of HcrB, HcrF, and Par. Color of the solid circles indicatesthe phenolic acid reductase the given sequence is homologous to, as follows: red, HcrB; blue, HcrF; and purple, HcrR. (B) Phylogenetic analysis of Pad. (C)Phylogenetic analysis of VprA.




http://aem.asm

.org/D

ownloaded from


vinylguaicol, suggesting a differential regulation of the enzyme and/or higher speci-ficity toward vinylphenol and vinylcatechol (20, 30). The L. rossiae FUA3583 Δpar1mutant did not reduce any of the substrates but decarboxylated ferulic acid to producevinyguaiacol, which was not detected in the wild-type strain. The metabolism ofhydroxycinnamic acids by the L. rossiae FUA3583 Δpar2 mutant was identical to that ofthe wild-type strain. The present study did not identify any of the genes coding forreduction of hydroxycinnamic acids in Weissella cibaria.

DISCUSSION

The pathway for hydroxycinnamic acid metabolism in lactic acid bacteria has beenknown for over a decade, with the first enzyme characterized being phenolic aciddecarboxylase in L. plantarum (16). The first hydroxycinnamic acid reductase was alsocharacterized in L. plantarum (21), but this phenolic acid reductase is absent in manyspecies that reduce hydroxycinnamates (this study). The genomic diversity observedbetween homofermentative and heterofermentative lactobacilli (28) indicates thatnovel enzymes may contribute to hydroxycinnamic acid metabolism in heterofermen-tative lactobacilli. This study identified several putative phenolic acid reductases inheterofermentative lactobacilli. Evidence for the contribution of phenolic acid reduc-tases to conversion of hydroxycinnamic acids was based on (i) characterization of thestrain-specific metabolism of hydroxycinnamic acids, (ii) comparative genomic analysesto identify putative phenolic acid reductases, (iii) identification of those putativeenzymes that were overexpressed in the presence of their substrates and (iv) charac-terization of the impact of deletion of Pad1 or Pad2 in L. rossiae on conversion ofhydroxycinnamic acids.

HcrR is a regulator of phenolic acid reductase in L. plantarum (21). An HcrR homologor ParR, a LysR-type transcriptional regulator identified in L. rossiae, was present in alllactobacilli that reduced hydroxycinnamic acids, suggesting that this metabolism isgenerally regulated (21). An HcrR or ParR homolog was not identified in W. cibaria 10M,but the lack of any recognizable genes for phenolic acid metabolism in that strainprevents any further conclusions related to Weissella.

The hcrA homologs present in heterofermentative hydroxycinnamic acid reducerswere not upregulated upon induction, nor are these genes part of operons related tophenolic acid metabolism. They resemble the genes for previously characterized NADH-dependent flavin reductases in L. johnsonii that play a role in hydrogen peroxideproduction (31). Moreover, HcrA homologs were also identified in L. reuteri, which doesnot metabolize hydroxycinnamic acids. HcrA in L. plantarum WCFS1 does not possessenzymatic activity, and the absence of reductase activity of hcrA knockout mutantsmight be explained by prevention of induction of hcrB, which is cotranscribed andpresent downstream of hcrA (21).

TABLE 2 Summary of hydroxycinnamic acid metabolism of Lactobacillales strainsa

StrainSinapic acid,dihydrosinapic acid

Ferulic acid Caffeic acid

Dihydroferulic acid4-Vinylguaiacol(Pad) Dihydrocaffeic acid

4-Vinylcatechol(Pad)

4-Ethylcatechol(VprA)

L. plantarum TMW1.460 � (hcrB) � (hcrB) � � (hcrB) � �L. plantarum FUA3584 � (hcrB) � (hcrB) � � (hcrB) � �L. rossiae FUA3583 � (par1/par2) � (par1) � � (par1) � �L. rossiae FUA3583 Δpar1 � (par2) � � � � �L. rossiae FUA3583 Δpar2 � (par1) � (par1) � � (par1) � �L. rossiae FUA3509 � (par1) � (par1) � � (par1) � �L. fermentum FUA3589 � (hcrF) � (hcrF) � � (hcrF) � �L. hammesii DSM16381 � � � � � �L. brevis TMW1.465 � � � � � �L. reuteri DSM20016 � � � � � �L. kunkeei DSM12361 � � � � � �Weissella cibaria 10M � � � � � �aA plus sign indicates the presence of genotype with text representing the type of phenolic acid reductase present with overexpression under the influence of aspecific substrate. A minus sign indicates the absence of a genotype. Shaded and unshaded boxes represent presence and absence of phenotype, respectively.




http://aem.asm

.org/D

ownloaded from


Most lactobacilli, with exception of the two species of the L. brevis, harbored fccA (9,22–25). FccA is 30 to 34% identical to HcrB and particularly contains the same FADbinding domain. Its presence in reductase-negative strains L. reuteri DSM 20016 and L.kunkeei DSM 12361, lack of a significant upregulation upon induction with any of thesubstrates, and the high similarity to Shewanella frigidimarina FccA, which is active asfumarate reductase (29), makes an involvement in phenolic acid metabolism unlikely.

HcrF in L. fermentum was more than 60% identical to HcrB in L. plantarum but lackedthe NADH binding domain, suggesting that only FMN binding and FAD bindingdomains are essential for reduction of phenolic acids. The presence of hcrR upstreamof hcrF suggests that the hcrR LysR family transcriptional regulator is responsible fortranscriptional regulation of hcrF (20, 21). High amino acid identity of HcrF with HcrB inthe aligned region and significant overexpression indicate its likely role in hydroxycin-namic acid metabolism.

L. rossiae C5 and FUA3583 reduced all three hydroxycinnamic acids but lacked eitherhcrB or hcrF. Two genes coding for putative phenolic acid reductases, par1 and par2,were identified in the genomes of L. rossiae strains. Despite the low amino acid identity,the presence of a domain architecture similar to that of other phenolic acid reductasespoints toward their role in reduction of hydroxycinnamic acids in L. rossiae. In addition,both strains overexpressed par1 in the presence of hydroxycinnamic acids, while onlysinapic acid induced par2 expression only in L. rossiae FUA3583. The lack of reductaseactivity observed in the L. rossiae FUA3583 Δpar1 mutant confirms its contribution asa phenolic acid reductase active on hydroxycinnamic acids. Homologs of Par1 and Par2are also present in L. plantarum but do not contribute to reduction of hydroxycinnamicacids (21). This, with the lack of influence of the L. rossiae FUA3583 Δpar2 mutant on thephenotype, further indicates the presence of an unknown spectrum of related com-pounds which can be potentially metabolized by lactic acid bacteria.

W. cibaria 10M showed phenolic acid reductase activity, comparable to other strainsof W. cibaria that were evaluated with respect to phenolic acid metabolism but notgenome sequenced (23). Surprisingly, none of the three phenolic acid reductases werefound in the sequenced genome or upon BLAST search against all genome sequencedstrains of Weissella species in the NCBI database. These results indicate the presence ofunidentified phenolic acid reductases in Weissella species.

Hydroxycinnamic acid metabolism in lactobacilli is highly strain specific (22–25).Heterofermentative lactobacilli use hydroxycinnamic acids as external electron accep-tors (23); in addition, reduction or decarboxylation of phenolic acids reduces theirantimicrobial activities (9). Three pad genes have been previously characterized fromdifferent lactobacilli (16, 17, 32). The differences in their substrate specificity can berelated to minor amino acid differences in the C-terminal region (32). There is alsoevidence of other unidentified phenolic acid decarboxylases (33), suggesting differinggenotypes among strains and/or a differential regulation. The Clostridium 2-enoatereductases also show strain-specific substrate specificity for reduction of hydroxycin-namic acids (18, 19). In lactobacilli, hydroxycinnamic acids are reduced by threedifferent enzymes, HcrB, HcrF, and Par1, with different strains and species havingdifferent substrate specificities and/or regulation explaining their differential hydroxy-cinnamic acid metabolism (this study). Remarkably, Par1 contributes to the metabolismof hydroxycinnamic acids in L. rossiae (this study) but not in L. plantarum (21), anddeletion of Par2 did not alter hydroxycinnamic acid metabolism in L. rossiae. Thepresence of metabolic enzymes with unknown substrates suggests that the metabolictoolset of lactobacilli to metabolize plant secondary metabolites is more extensive thancurrently known.

To relate the ecological relevance of phenolic acid metabolism in lactobacilli, wescreened Lactobacillus genomes for the presence or absence of genes related tophenolic acid metabolism. Species belonging to 11 of 24 Lactobacillus phylogeneticgroups (28) were represented on the tree. Species of the L. plantarum group arecapable of metabolizing hydroxycinnamic acids, which can be attributed to their broadmetabolic potential that relates to the nomadic lifestyle (34). Hydroxycinnamic acid




http://aem.asm

.org/D

ownloaded from


metabolism was also frequently identified in species of the L. alimentarius group.Genomes of several species known to reduce hydroxycinnamic acids, including L.crispatus, L. collinoides, and L. hilgardii, contained par1/par2 sequences as sole genes forphenolic acid reductases (35–37). Heterofermentative lactobacilli harboring hcrF exclu-sively belonged to the L. reuteri and L. collinoides groups, while homofermentativespecies with hcrF belonged to the L. delbrueckii group, including L. gasseri. Some L.gasseri strains also displayed reductase activity in the absence of hcrB (38). L. reuterigroup organisms and L. delbrueckii group organisms share the same ecological niche inthe intestine of humans and animals (34, 39–41); the almost exclusive presence of HrcFin these species may reflect a lifestyle adaptation. With the exception of L. curieae andL. vaccinostercus, the 16 species harboring HcrB were all homofermentative.

Metabolism of hydroxycinnamic acids among heterofermentative lactobacilli wasmost frequent in species with a free-living lifestyle, particularly in organisms of the L.collinoides and L. buchneri groups. These organisms were mostly isolated from plant orfermented plant products, where they regularly encounter phenolic acids (34, 42). Noneof the heterofermentative insect-adapted lactobacilli possessed genes related to phe-nolic acid metabolism. Vertebrate-adapted lactobacilli of the L. reuteri, L. delbrueckii,and L. salivarius groups also have a significant presence on the three trees, which mayrelate to their interaction with phenolic acids passing through the gut due to con-sumption of foods and dietary fibers rich in phenolic compounds by the host (4). Sevenof the 13 species that belong to the L. delbrueckii group harbor vrpA but not pad andare species with an insect-adapted lifestyle. Phenolic acid decarboxylase and vinylreductase are thought to be part of the same metabolic pathway (20), and the presenceof only VrpA may point to syntrophic interactions among insect gut microbiota.Alternatively, VrpA homologs in insect-adapted lactobacilli may be active on substratesthat relate to the intestine of bees but are not related to vinyl derivatives of hydroxy-cinnamic acids (43).

In conclusion, the current study expanded knowledge on the genetic determinantsof the diverse metabolism of hydroxycinnamic acids by lactic acid bacteria that explainspecies- and strain-specific metabolic differences. Two novel phenolic acid reductaseswere identified, with Par1 being active in L. rossiae and HcrF being likely responsible inL. fermentum for reduction of hydroxycinnamic acids. The phenotypic analysis of strainssuggests that the genotype, the substrate specificity of the enzymes, and the regulationof gene expression are responsible for the strain-specific differences in metabolism. Ouranalyses also provide evidence for additional, yet-uncharacterized phenolic acid reduc-tases in Weissella spp. In addition, Par1/Par2 enzymes with homologies to phenolic acidreductases in L. plantarum along with par2 in L. rossiae appear to be inactive onhydroxycinnamic acids and thus may convert related secondary metabolites of plantsor fungi expanding the utility of lactic acid bacteria.

MATERIALS AND METHODSBacterial strains and growth conditions. Lactobacillus plantarum TMW1.460 (44) and Lactobacillus

brevis TMW1.465 (45) isolated from spoiled beer, Lactobacillus hammesii DSM16381 (46) and Lactobacillusrossiae C5 isolated from sourdough (24), L. rossiae FUA3583, L. plantarum FUA3584, and L. fermentumFUA3589 isolated from Mahewu (47) as well as Lactobacillus reuteri DSM20016, Lactobacillus kunkeeiDSM12361, and Weissella cibaria 10M (48) were used in this study. Strains were subcultured from – 80°Cstock and grown in modified De Man, Rogosa and Sharpe (mMRS) medium (49) at 30°C undermicroaerophilic conditions.

Chemicals. Ferulic acid and caffeic acid were obtained from Extrasynthèse (Genay, France). Sinapicacid, dihydrosinapic acid, dihydrocaffeic acid, 4-vinylguaiacol, 4-ethylguaiacol, and 4-ethylcatechol wereobtained from Sigma-Aldrich (ON, Canada). Dihydroferulic acid was purchased from MP Biomedicals andcomponents for mMRS media were purchased from BD (ON, Canada) and Sigma-Aldrich.

Hydroxycinnamic acid metabolism of lactic aid bacterial strains. Sinapic acid, ferulic acid, orcaffeic acid was added to mMRS medium, and supplemented media were inoculated with 10% overnightculture of bacterial strains and incubated for 24 h at 30°C (3). Uninoculated mMRS media containingcorresponding hydroxycinnamic acids were used as controls. Following the incubation, cells wereremoved by centrifugation; the supernatant was acidified to pH 1.5 using HCl and extracted twice withethyl acetate (9). The extracts were combined and analyzed with an Agilent 1200 series HPLC systemequipped with an Agilent Eclipse XDB C18 column (4.6 by 150 mm; 5 �m) coupled to an UV detector.Ten-microlitersamples were injected and eluted at a flow rate of 0.7 ml/min using mobile phases




http://aem.asm

.org/D

ownloaded from


consisting of 0.1% (vol/vol) formic acid in water (mobile phase A) and 0.1% formic acid in water-acetonitrile (10:90, vol/vol) (mobile phase B). The gradient applied on phase B was as follows: 10% to 15%(0 to 6 min), 15% to 100% (6 to 14 min), isocratic at 100% (21 to 24 min), and 100% to 10% (24 to 30 min).Quantification was performed using external standards at 280 nm, with duplicate independent experimentsperformed. Due to the low stability of vinylcatechol, relative peak areas were calculated as a ratio of peak areaout of 1, with vinylcatechol peak area divided by total peak area subtracted by interference peak area.

Comparative genomics and sequence analysis. Genome sequences of L. hammesii DSM16381, L.brevis TMW1.465, L. reuteri DSM20016, L. kunkeei DSM12361, and L. fermentum FUA3589 were retrievedfrom NCBI. Whole-genome shotgun sequences of L. plantarum TMW1.460, L. rossiae C5, W. cibaria 10M,L. rossiae FUA3583, and L. plantarum FUA3584 were obtained in this study. DNA was isolated using aWizard genomic DNA purification kit (Promega, WI). The quantity and quality of DNA were verified by gelelectrophoresis and with a NanoDrop One spectrophotometer (Thermo Scientific, DE). Sequencing wasperformed using Illumina TruSeq on a HiSeq2500 platform with high-output run mode by McGillUniversity and Génome Québec Innovation Centre (Montreal, QC, Canada). The quality check of 125-bppaired-end reads was done using the FastQC tool (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Sequence assembly was performed using SPAdes (50) and MeDuSa (51). Genomes were thenannotated using the RAST server (52).

TABLE 3 qPCR primer sequences

Strain(s) Primera Primer sequence (5=¡3=)L. plantarum TMW1.460 and

L. plantarum FUA3584hcrR-F/R CGGTGAGCTGGACTTCTTAAT/GTTGACGGTGTTCGGGATAAhcrA-F/R GGGACGAATGCAACCAAATC/TCGGTCTTCCGGTTCATTAAAhcrB-F/R CGCATACCTGACTGCCAATA/CAGTCCGTTGACCACCTAAAhcrC-F/R TGGATCACCGACTTTCATCTTC/TGGATCACCGACTTTCATCTTCparR-F/R GCATGCAACCCGCAATTATC/ATCCGTCAAATCAGCCAAGAAhcrA2-F/R TGGCGGATAAGATCGAACAAG/AGGGCTGGATATGGAATGATAACpar1-F/R CAGTATCAAGGTGGCGGTAAT/CTAGTATTCGCATCCGGCTTAGpar2-F/R GTGAACTGGCACGGTAACT/CTAGTACGTGGGCACCATTATCfccA-F/R CTGGTGCCTACTTGATCTTTGA/CCAAGTCAGTCCCAGTCTTTACpadR-F/R TGAAGCGACTAGAAGAACAAGG/GCGGTGATGTGGTAGAGTTTpad-F/R CATGTTGACCGAAGGCATTTAC/CGTACCGTGTAGTTTCTTCTCATvrpA-F/R ACCGGTGGTTACCTCAATAATC/ACCGTCACCAGTTCCTTTACpfk (housekeeping)-F/R CAATTACGGCTTTGCTGGATTAG/CTGGATAACGTGCGGAGTATAG

L. rossiae C5 parR-F/R ATTGGTGCCGAGGATGTAAG/TTCGCTGAAAGCCGGAAThcrA2-F/R GATGGGTTGTTTGCCGATTTC/CTTGCTTCTTGGCGATTGATGhcrA3-F/R ATGCTGCCTGCCGATAAA/CCAGTCCGTAACATGTGGATAApar1-F/R ATCGTGGTTCGGGTAACTTATG/GAATCTCCTCCACGGCTTTACpar2-F/R GCCAAACAGGGCAGAGATTAT/GTGGACCGCCTTGTGTATTTfccA-F/R CCGGACCATTCTTCGCTATT/CTAGGATGATCTGACCGCTTTCphosphoketolase (housekeeping)-F/R GAGCGATGCTGACTTGACTAA/CCATGTCTTGATGAGCCTTCT

L. rossiae FUA3583 parR-F/R GTTACCACTACACAGGCGTTTA/CGGTGAAATGCTGGCAATAAGhcrA2-F/R CGATGACGACGGAAATTTGTTAG/CTTGCTTCTTGGCGATTGATGhcrA3-F/R CGATGCGTGACAGTGAGAATA/CTAACTTCTGACCGGTCTTAGCpar1-F/R GCGGATGATGGTAGTCGTTATG/AGATGAGGATGCGCGTAAACpar2-F/R GGAATACGAAGGTGTCCATTCA/TTATCAGCGCGGTCCATAACfccA-F/R AGACACAGGTATTGGACGAAAG/AGCCTGACGACCAAAGATTACpadR-F/R GATGTTGACTGATGGCTGGA/TCGTGACCATCTGCTGTAATCpad-F/R GGGTTGAGGAACATCCAGAAA/GCAAATTCTGGCACCACTAACphosphoketolase (housekeeping)-F/R GAGCGATGCTGACTTGACTAA/TGTCTTGGTGAGCCTTCTTG

L. fermentum FUA3589 hcrR-F/R CAAGGTGGTGATGGTCTCAA/GTTGATCGGCGTTTCCTTAATChcrA2-F/R CGATCGAATGGCTGTCCTATAA/GAAGAACCCTGGGTGAAGTAAGhcrA3-F/R ATCATGCCATCATCCGAATACA/GTGTCGAAGTCGGCGAATAAhcrF-F/R GTGACCGTCCGTGCTAAAT/TCGTCAATGTGCTCCCAATAGfccA-F/R CAGTGCTAAGGCAGTGATTCT/GGTTGGTTGGTCGTCTTGTApadR-F/R CCTACGTCATCTTAGGGATCATTG/TTGGCTGTGCGAGGATTTpad-F/R GGGAATACGAATGGTACGCTAAA/TGTGGGCTTCTTGGTCATTCphosphoketolase (housekeeping)-F/R TGGCTGCTTCATGGTTCTC/CGGGAAAGGATAGTTGGGTTAG

L. reuteri DSM20016 hcrA2-F/R CCTGAGGCTGTTGCTGATATT/CAGAGGTAACAGAGTGGTTGTATThcrA3-F/R GCGGCCTTAAAGAGTACAGTAG/GAAGCTTGGCGTCCATAAGAfccA-F/R GTGTTCTGGAAGGGCAGTAATC/GTGGTGAAGGTGCAATCTTAGTphosphoketolase(housekeeping)-F/R CAGAACACCAAGCTGAAGGA/GAATCAACAACACGACCGAATG

L. kunkeei DSM12361 fccA-F/R GGTCGAGTTACCACCTAACTTATC/GTTCTGGTGCCACAGGATTAphosphoketolase (housekeeping)-F/R TCAGCAAACACACCGAATAGA/ACTGGTTAGGTGCCGTTATG

aF, forward; R, reverse.




http://aem.asm

.org/D

ownloaded from

https://www.bioinformatics.babraham.ac.uk/projects/fastqc/https://www.bioinformatics.babraham.ac.uk/projects/fastqc/https://aem.asm.orghttp://aem.asm.org/

Genes involved in the hydroxycinnamate reductase operon of L. plantarum WCFS1 (21) were used asquery sequences to search for homologous protein sequences in the sample genomes using BLAST�.Protein sequence analysis was performed using InterProScan (53) and InterPro tools (54).

RNA isolation and quantification of relative gene expression by RT-qPCR. mMRS media con-taining 1 mM sinapic acid, caffeic acid, and ferulic acid were inoculated with overnight cultures oflactobacilli. mMRS medium without addition of hydroxycinnamic acids was used as a reference conditionfor each strain. Strains were grown to early exponential phase corresponding to an optical density at 600nm (OD600) of 0.3 to 0.4. RNA was stabilized by adding 2 volumes of RNAprotect bacterial reagent(Qiagen, ON, Canada) to 1 volume of bacterial cultures. RNA was isolated with the RNeasy minikit(Qiagen) according to the manufacturer’s instructions. RNA quantification and purity were measuredusing a NanoDrop One spectrophotometer (Thermo Scientific, DE), and DNA was hydrolyzed with RQ1RNase�free DNase (Promega) according to the manufacturer’s instructions. cDNA was synthesized withthe QuantiTect reverse transcription kit (Qiagen); 1 �g of RNA was used as the template, and reversetranscription was done according to the manufacturer’s instructions.

qPCR was performed in a 7500 fast real�time PCR instrument (Applied Biosystems, Life Technologies,ON, Canada) with a QuantiFast SYBR green PCR kit (Qiagen). Primers were designed using the IntegratedDNA Technologies (IDT) PrimerQuest Tool (Table 3). Gene expression relative to the reference conditionswas calculated according to the method by Pfaffl (55). The phosphoketolase gene was used as thehousekeeping gene for heterofermentative lactobacilli, and the phosphofructokinase gene was used asthe housekeeping gene for strains of L. plantarum (56). mRNA abundance under experimental conditionswere normalized to the mRNA abundance under reference conditions, i.e., during growth of lactobacilliin the absence of phenolic acids. Triplicate independent experiments were conducted (replicate bacterialcultures), and statistical analysis was performed using one-way analysis of variance (ANOVA) withHolm-Sidak post hoc analysis.

Construction of L. rossiae �par1 and �par2 mutants. Upstream and downstream flanking regionsof target genes (800 to 1,000 bp) were PCR amplified and cloned into counterselection plasmid pVPL3002(57) using the ligation cycling reaction (LCR) (58). Primers for mutant construction are shown in Table 4.The resulting recombinant plasmids, pVPL3002/Δpar1 and pVPL3002/Δpar2, were transformed intoEscherichia coli EC1000 (59) and plated on Luria-Bertani (LB) medium containing 300 mg/liter of eryth-romycin for selection. Electrocompetent L. rossiae FUA3583 cells were first transformed at 2.5 kV, 25 �F,and 400 � with 4 �g of plasmid pVE6007 (60). After 3 h of recovery, cells were plated in the presence of5 mg/liter of chloramphenicol for selection of transformants. L. rossiae FUA3583 harboring pVE6007 wasthen transformed with 2 to 3 �g of plasmid DNA (pVPL3002/Δpar1 or pVPL3002/Δpar2) using the sameconditions as mentioned above. After recovery in medium with 5 mg/liter of chloramphenicol, afast-track genome editing approach was followed (57). Briefly, the recovered cells were transferred into40 ml of medium containing 2.5 mg/liter of erythromycin and 5 mg/liter of chloramphenicol. Cells werewashed with medium not containing any antibiotics after 48 to 60 h of incubation. Cells were thensubcultured 2 or 3 times in medium containing 2.5 mg/liter of erythromycin. Cells were once again

TABLE 4 Primers for mutant construction

Primer Description Primer sequence (5=¡3=)oVPL 188 F (57) Amplifies pVPL3002 backbone ATCCTCTAGAGTCGACCTGCoVPL 187 R (57) TACCGAGCTCGAATTCACTGGpar1 U/S F Upstream flanking region of par1 in L. rossiae FUA3583 GCAGCCAGATAGCCTGAAACpar1 U/S R CGACTGGCAGTTGCGCCAGCTGCGCpar1 D/S F Downstream flanking region of par1 in L. rossiae FUA3583 AAGACGTTGGTCGTAAGGCCGTGpar1 D/S R CATAGCGGCAGTGAACTTGApar1 BO1 LCR bridging oligonucleotide for pVPL3002/Δpar1 AAACGACGGCCAGTGAATTCGAGCTCGGTAGCAGCCAG

ATAGCCTGAAACAATTCGTTGGpar1 BO2 LCR bridging oligonucleotide for pVPL3002/Δpar1 CTTTGGCGCAGCTGGCGCAACTGCCAGTCGAAGACGTT

GGTCGTAAGGCCGTGGAGGAGApar1 BO3 LCR bridging oligonucleotide for pVPL3002/Δpar1 GAATCCTTCATCAAGTTCACTGCCGCTATGATCCTCTA

GAGTCGACCTGCAGGCATGCAApar1 DCO F DCO screening for Δpar1 in L. rossiae FUA3583 AATCGTTGATCCGGCATTACpar1 DCO R TCACACGCGATAGGTCTGAGpar2 U/S F Upstream flanking region of par2 in L. rossiae FUA3583 ACGCATGGTCTACCAGTTCCpar2 U/S R TAACGGGTGTTACCACCTTCATGpar2 D/S F Downstream flanking region of par2 in L. rossiae FUA3583 ATCAGTATCTAGCCGCGCTATTpar2 D/S R GCAGTTGTCAGCAAGGAACApar2 BO1 LCR bridging oligonucleotide for pVPL3002/Δpar2 AAACGACGGCCAGTGAATTCGAGCTCGGTAACGCATGG

TCTACCAGTTCCTGAAACCGTGpar2 BO2 LCR bridging oligonucleotide for pVPL3002/Δpar2 AAACGGACATGAAGGTGGTAACACCCGTTAATCAGTATC

TAGCCGCGCTATTAAAGACGCpar2 BO3 LCR bridging oligonucleotide for pVPL3002/Δpar2 TGCCAACGGATGTTCCTTGCTGACAACTGCATCCTCTAG

AGTCGACCTGCAGGCATGCAApar2 DCO F DCO screening for Δpar2 in L. rossiae FUA3583 GATTCCAATCGCCATAATGCpar2 DCO R CCATTAATTGCAGGCCAGTT




http://aem.asm

.org/D

ownloaded from


washed and subcultured into medium without any antibiotics for one passage, followed by plating onmedium containing 500 mg/liter of vancomycin for selection of double-crossover (DCO) mutants. L.rossiae was grown in MRS medium plus cysteine (0.5 g/liter) at 37°C under anaerobic conditions duringmutant construction. Mutants with deletions in par1 and par2 were confirmed using colony PCR, whilethe phenotype was tested using HPLC.

Phylogenetic analysis. HcrB, Par1, and HcrF protein sequences were used as query sequences tosearch for homologs using BLAST on NCBI with default parameters across all species of Lactobacillus andPediococcus for which genome sequences were available in July 2018. Sequences with greater than 80%query cover and 40% amino acid identity were retrieved for each species. Sequences differing by morethan 10% in amino acid identity with respect to the query sequence within a species were also retrieved.Multiple-sequence alignment was performed using MUSCLE (61). A maximum likelihood tree was constructedusing IQ TREE web server (62) using the best fit model predicted by ModelFinder (63). Bootstrap values for1,000 replicates were calculated with ultrafast bootstrap approximation (UFBoot) (64). Phylogenetic analysisfor phenolic acid decarboxylase and vinylphenol reductase was also performed using the same process withPad (YP_004891133.1) and VprA (YP_004890680.1) genes from L. plantarum WCFS1 used as querysequences for respective tree constructions. Tree visualization was done using the iTOL online tool (65).

Data availability. The NCBI genome accession numbers for strains studied are as follows:NZ_AZFS00000000.1 (L. hammesii DSM 16381), GCA_000833395.1 (L. brevis TMW1.465), NC_009513.1 (L.reuteri DMS20016), NZ_AZCK00000000.1 (L. kunkeei DSM12361), NZ_SMZH00000000.1 (L. fermentumFUA3589), WEZQ00000000 (L. rossiae C5), WEZT00000000 (L. rossiae FUA3583), WEZR00000000 (L.plantarum TMW1.460), WEZU00000000 (L. plantarum FUA3584), and WEZS00000000 (W. cibaria 10M).

ACKNOWLEDGMENTSWe thank Felicitas Pswarayi and Danielle Balay for assistance in genome sequencing

of strains.

REFERENCES1. Andreasen MF, Christensen LP, Meyer AS, Hansen Å. 2000. Content of

phenolic acids and ferulic acid dehydrodimers in 17 rye (Secale cereale L.)varieties. J Agric Food Chem 48:2837–2842. https://doi.org/10.1021/jf991266w.

2. Dabrowski KJ, Sosulski FW. 1984. Composition of free and hydrolyzablephenolic acids in defatted flours of ten oilseeds. J Agric Food Chem32:128 –130. https://doi.org/10.1021/jf00121a032.

3. Svensson L, Sekwati-Monang B, Lutz DL, Schieber R, Gänzle MG. 2010.Phenolic acids and flavonoids in nonfermented and fermented redsorghum (Sorghum bicolor (L.) Moench). J Agric Food Chem 58:9214 –9220. https://doi.org/10.1021/jf101504v.

4. Acosta-Estrada BA, Gutiérrez-Uribe JA, Serna-Saldívar SO. 2014. Boundphenolics in foods, a review. Food Chem 152:46 –55. https://doi.org/10.1016/j.foodchem.2013.11.093.

5. Kim Y, Keogh JB, Clifton PM, Kim Y, Keogh J, Clifton P. 2016. Polyphenolsand glycemic control. Nutrients 8:17. https://doi.org/10.3390/nu8010017.

6. Shahidi F, Yeo J, Shahidi F, Yeo J. 2018. Bioactivities of phenolics byfocusing on suppression of chronic diseases: a review. Int J Mol Sci19:1573. https://doi.org/10.3390/ijms19061573.

7. Chao C, Mong M, Chan K, Yin M. 2010. Anti-glycative and anti-inflammatory effects of caffeic acid and ellagic acid in kidney of diabeticmice. Mol Nutr Food Res 54:388 –395. https://doi.org/10.1002/mnfr.200900087.

8. Janicke B, Hegardt C, Krogh M, Önning G, Åkesson B, Cirenajwis HM,Oredsson SM. 2011. The antiproliferative effect of dietary fiber phenoliccompounds ferulic acid and p-coumaric acid on the cell cycle of Caco-2cells. Nutr Cancer 63:611– 622. https://doi.org/10.1080/01635581.2011.538486.

9. Sánchez-Maldonado AF, Schieber A, Gänzle MG. 2011. Structure-functionrelationships of the antibacterial activity of phenolic acids and theirmetabolism by lactic acid bacteria. J Appl Microbiol 111:1176 –1184.https://doi.org/10.1111/j.1365-2672.2011.05141.x.

10. Sánchez Maldonado AF, Mudge E, Gänzle MG, Schieber A. 2014. Extrac-tion and fractionation of phenolic acids and glycoalkaloids from potatopeels using acidified water/ethanol-based solvents. Food Res Int 65:27–34. https://doi.org/10.1016/j.foodres.2014.06.018.

11. Bai Y, Findlay B, Sanchez Maldonado AF, Schieber A, Vederas JC, GänzleMG. 2014. Novel pyrano and vinylphenol adducts of deoxyanthocyani-dins in sorghum sourdough. J Agric Food Chem 62:11536 –11546.https://doi.org/10.1021/jf503330b.

12. Azevedo J, Oliveira J, Cruz L, Teixeira N, Brás NF, De Freitas V, MateusN. 2014. Antioxidant features of red wine pyranoanthocyanins: ex-

perimental and theoretical approaches. J Agric Food Chem 62:7002–7009. https://doi.org/10.1021/jf404735j.

13. Lai KK, Lorca GL, Gonzalez CF. 2009. Biochemical properties of twocinnamoyl esterases purified from a Lactobacillus johnsonii strain iso-lated from stool samples of diabetes-resistant rats. Appl Environ Micro-biol 75:5018 –5024. https://doi.org/10.1128/AEM.02837-08.

14. Esteban-Torres M, Reverón I, Mancheño JM, de las Rivas B, Muñoz R.2013. Characterization of a feruloyl esterase from Lactobacillus planta-rum. Appl Environ Microbiol 79:5130 –5136. https://doi.org/10.1128/AEM.01523-13.

15. Esteban-Torres M, Landete JM, Reverón I, Santamaría L, de las Rivas B,Muñoz R. 2015. A Lactobacillus plantarum esterase active on a broadrange of phenolic esters. Appl Environ Microbiol 81:3235–3242. https://doi.org/10.1128/AEM.00323-15.

16. Cavin JF, Barthelmebs L, Diviès C. 1997. Molecular characterization of aninducible p-coumaric acid decarboxylase from Lactobacillus plantarum:gene cloning, transcriptional analysis, overexpression in Escherichia coli,purification, and characterization. Appl Environ Microbiol 63:1939 –1944.

17. Rodríguez H, Landete JM, Curiel JA, de las Rivas B, Mancheño JM, MuñozR. 2008. Characterization of the p-coumaric acid decarboxylase fromLactobacillus plantarum CECT 748T. J Agric Food Chem 56:3068 –3072.https://doi.org/10.1021/jf703779s.

18. Mordaka PM, Hall SJ, Minton N, Stephens G. 2018. Recombinant expres-sion and characterisation of the oxygen-sensitive 2-enoate reductasefrom Clostridium sporogenes. Microbiology 164:122–132. https://doi.org/10.1099/mic.0.000568.

19. Sun J, Lin Y, Shen X, Jain R, Sun X, Yuan Q, Yan Y. 2016. Aerobicbiosynthesis of hydrocinnamic acids in Escherichia coli with a strictlyoxygen-sensitive enoate reductase. Metab Eng 35:75– 82. https://doi.org/10.1016/j.ymben.2016.02.002.

20. Santamaría L, Reverón I, de Felipe FL, de las Rivas B, Muñoz R. 2018.Ethylphenol formation by Lactobacillus plantarum: identification of theenzyme involved in the reduction of vinylphenols. Appl Environ Microbiol84:e01064-18. https://doi.org/10.1128/AEM.01064-18.

21. Santamaría L, Reverón I, de Felipe FL, de las Rivas B, Muñoz R. 2018.Unravelling the reduction pathway as an alternative metabolic route tohydroxycinnamate decarboxylation in Lactobacillus plantarum. Appl En-viron Microbiol 84:e01123-18. https://doi.org/10.1128/AEM.01123-18.

22. Filannino P, Bai Y, Di Cagno R, Gobbetti M, Gänzle MG. 2015. Metabolismof phenolic compounds by Lactobacillus spp. during fermentation ofcherry juice and broccoli puree. Food Microbiol 46:272–279. https://doi.org/10.1016/j.fm.2014.08.018.

23. Filannino P, Gobbetti M, De Angelis M, Di Cagno R. 2014. Hydroxycin-




http://aem.asm

.org/D

ownloaded from

https://www.ncbi.nlm.nih.gov/protein/YP_004891133.1https://www.ncbi.nlm.nih.gov/protein/YP_004890680.1https://www.ncbi.nlm.nih.gov/nuccore/NZ_AZFS00000000.1https://www.ncbi.nlm.nih.gov/genome/1110?genome_assembly_id=219099https://www.ncbi.nlm.nih.gov/nuccore/NC_009513.1https://www.ncbi.nlm.nih.gov/nuccore/NZ_AZCK00000000.1https://www.ncbi.nlm.nih.gov/nuccore/NZ_SMZH00000000.1https://www.ncbi.nlm.nih.gov/nuccore/WEZQ00000000https://www.ncbi.nlm.nih.gov/nuccore/WEZT00000000https://www.ncbi.nlm.nih.gov/nuccore/WEZR00000000https://www.ncbi.nlm.nih.gov/nuccore/WEZU00000000https://www.ncbi.nlm.nih.gov/nuccore/WEZS00000000https://doi.org/10.1021/jf991266whttps://doi.org/10.1021/jf991266whttps://doi.org/10.1021/jf00121a032https://doi.org/10.1021/jf101504vhttps://doi.org/10.1016/j.foodchem.2013.11.093https://doi.org/10.1016/j.foodchem.2013.11.093https://doi.org/10.3390/nu8010017https://doi.org/10.3390/ijms19061573https://doi.org/10.1002/mnfr.200900087https://doi.org/10.1002/mnfr.200900087https://doi.org/10.1080/01635581.2011.538486https://doi.org/10.1080/01635581.2011.538486https://doi.org/10.1111/j.1365-2672.2011.05141.xhttps://doi.org/10.1016/j.foodres.2014.06.018https://doi.org/10.1021/jf503330bhttps://doi.org/10.1021/jf404735jhttps://doi.org/10.1128/AEM.02837-08https://doi.org/10.1128/AEM.01523-13https://doi.org/10.1128/AEM.01523-13https://doi.org/10.1128/AEM.00323-15https://doi.org/10.1128/AEM.00323-15https://doi.org/10.1021/jf703779shttps://doi.org/10.1099/mic.0.000568https://doi.org/10.1099/mic.0.000568https://doi.org/10.1016/j.ymben.2016.02.002https://doi.org/10.1016/j.ymben.2016.02.002https://doi.org/10.1128/AEM.01064-18https://doi.org/10.1128/AEM.01123-18https://doi.org/10.1016/j.fm.2014.08.018https://doi.org/10.1016/j.fm.2014.08.018https://aem.asm.orghttp://aem.asm.org/

namic acids used as external acceptors of electrons: an energetic ad-vantage for strictly heterofermentative lactic acid bacteria. Appl EnvironMicrobiol 80:7574 –7582. https://doi.org/10.1128/AEM.02413-14.

24. Ripari V, Bai Y, Gänzle MG. 2019. Metabolism of phenolic acids in wholewheat and rye malt sourdoughs. Food Microbiol 77:43–51. https://doi.org/10.1016/j.fm.2018.08.009.

25. Filannino P, Di Cagno R, Addante R, Pontonio E, Gobbetti M. 2016.Metabolism of fructophilic lactic acid bacteria isolated from the Apismellifera L. bee gut: phenolic acids as external electron acceptors.Appl Environ Microbiol 82:6899 – 6911. https://doi.org/10.1128/AEM.02194-16.

26. Muñoz R, de las Rivas B, López de Felipe F, Reverón I, Santamaría L,Esteban-Torres M, Curiel JA, Rodríguez H, Landete JM. 2017. Biotrans-formation of phenolics by Lactobacillus plantarum in fermented foods, p63– 83. In Frias J, Martinez-Villaluenga C, Peñas E (ed), Fermented foodsin health and disease prevention. Academic Press, London, United King-dom.

27. Gänzle MG. 2015. Lactic metabolism revisited: metabolism of lactic acidbacteria in food fermentations and food spoilage. Curr Opin Food Sci2:106 –117. https://doi.org/10.1016/j.cofs.2015.03.001.

28. Zheng J, Ruan L, Sun M, Gänzle MG. 2015. A genomic view of lactobacilliand pediococci demonstrates that phylogeny matches ecology andphysiology. Appl Environ Microbiol 81:7233–7243. https://doi.org/10.1128/AEM.02116-15.

29. Pealing SL, Lysek DA, Taylor P, Alexeev D, Reid GA, Chapman SK,Walkinshaw MD. 1999. Crystallization and preliminary X-ray analysis offlavocytochrome c3, the fumarate reductase from Shewanella frigidima-rina. J Struct Biol 127:76 –78. https://doi.org/10.1006/jsbi.1999.4139.

30. Reverón I, de las Rivas B, Muñoz R, López de Felipe F. 2012. Genome-wide transcriptomic responses of a human isolate of Lactobacillus plan-tarum exposed to p-coumaric acid stress. Mol Nutr Food Res 56:1848 –1859. https://doi.org/10.1002/mnfr.201200384.

31. Hertzberger R, Arents J, Dekker HL, Pridmore RD, Gysler C, KleerebezemM, de Mattos M. 2014. H2O2 production in species of the Lactobacillusacidophilus group: a central role for a novel NADH-dependent flavinreductase. Appl Environ Microbiol 80:2229 –2239. https://doi.org/10.1128/AEM.04272-13.

32. Landete JM, Rodríguez H, Curiel JA, de las Rivas B, Mancheño JM, MuñozR. 2010. Gene cloning, expression, and characterization of phenolic aciddecarboxylase from Lactobacillus brevis RM84. J Ind Microbiol Biotechnol37:617– 624. https://doi.org/10.1007/s10295-010-0709-6.

33. Barthelmebs L, Divies C, Cavin J-F. 2000. Knockout of the p-coumaratedecarboxylase gene from Lactobacillus plantarum reveals the existenceof two other inducible enzymatic activities involved in phenolic acidmetabolism. Appl Environ Microbiol 66:3368 –3375. https://doi.org/10.1128/aem.66.8.3368-3375.2000.

34. Duar RM, Lin XB, Zheng J, Martino ME, Grenier T, Pérez-Muñoz ME,Leulier F, Gänzle M, Walter J. 2017. Lifestyles in transition: evolution andnatural history of the genus Lactobacillus. FEMS Microbiol Rev 41:S27–S48. https://doi.org/10.1093/femsre/fux030.

35. Buron N, Coton M, Desmarais C, Ledauphin J, Guichard H, Barillier D,Coton E. 2011. Screening of representative cider yeasts and bacteriafor volatile phenol-production ability. Food Microbiol 28:1243–1251.https://doi.org/10.1016/j.fm.2011.05.001.

36. van Beek S, Priest FG. 2000. Decarboxylation of substituted cinnamicacids by lactic acid bacteria isolated during malt whisky fermentation.Appl Environ Microbiol 66:5322–5328. https://doi.org/10.1128/aem.66.12.5322-5328.2000.

37. de las Rivas B, Rodríguez H, Curiel JA, Landete JM, Muñoz R. 2009.Molecular screening of wine lactic acid bacteria degrading hydroxycin-namic acids. J Agric Food Chem 57:490 – 494. https://doi.org/10.1021/jf803016p.

38. Oh NS, Lee JY, Oh S, Joung JY, Kim SG, Shin YK, Lee K-W, Kim SH, Kim Y.2016. Improved functionality of fermented milk is mediated by thesynbiotic interaction between Cudrania tricuspidata leaf extract andLactobacillus gasseri strains. Appl Microbiol Biotechnol 100:5919 –5932.https://doi.org/10.1007/s00253-016-7414-y.

39. Tannock GW, Wilson CM, Loach D, Cook GM, Eason J, O’Toole PW,Holtrop G, Lawley B. 2012. Resource partitioning in relation to cohabi-tation of Lactobacillus species in the mouse forestomach. ISME J6:927–938. https://doi.org/10.1038/ismej.2011.161.

40. Lin XB, Wang T, Stothard P, Corander J, Wang J, Baines JF, Knowles SCL,Baltru�naitė L, Tasseva G, Schmaltz R, Tollenaar S, Cody LA, Grenier T, WuW, Ramer-Tait AE, Walter J. 2018. The evolution of ecological facilitation

within mixed-species biofilms in the mouse gastrointestinal tract. ISME J12:2770 –2784. https://doi.org/10.1038/s41396-018-0211-0.

41. Leser TTD, Amenuvor JZJ, Jensen TTK, Lindecrona RRH, Boye M, Møller K.2002. Culture-independent analysis of gut bacteria: the pig gastrointes-tinal tract microbiota revisited. Appl Environ Microbiol 68:673– 690.https://doi.org/10.1128/aem.68.2.673-690.2002.

42. Gänzle MG. 2019. Fermented foods, p 855–900. In Doyle MP, Diez-Gonzalez F, Hill C (ed), Food microbiology: fundamentals and frontiers,5th ed. ASM Press, Washington, DC.

43. Kwong WK, Moran NA. 2015. Evolution of host specialization in gutmicrobes: the bee gut as a model. Gut Microbes 6:214 –220. https://doi.org/10.1080/19490976.2015.1047129.

44. Ulmer HM, Ganzle MG, Vogel RF. 2000. Effects of high pressure onsurvival and metabolic activity of Lactobacillus plantarum TMW1.460.Appl Environ Microbiol 66:3966 –3973. https://doi.org/10.1128/aem.66.9.3966-3973.2000.

45. Behr J, Gänzle MG, Vogel RF. 2006. Characterization of a highly hop-resistant Lactobacillus brevis strain lacking hop transport. Appl EnvironMicrobiol 72:6483– 6492. https://doi.org/10.1128/AEM.00668-06.

46. Valcheva R, Korakli M, Onno B, Prévost H, Ivanova I, Ehrmann MA,Dousset X, Gänzle MG, Vogel RF. 2005. Lactobacillus hammesii sp. nov.,isolated from French sourdough. Int J Syst Evol Microbiol 55:763–767.https://doi.org/10.1099/ijs.0.63311-0.

47. Pswarayi F, Gänzle MG, Pswarayi F, Gänzle MG. 2019. Composition andorigin of the fermentation microbiota of mahewu, a Zimbabwean fer-mented cereal beverage. Appl Environ Microbiol 85:e03130-18. https://doi.org/10.1128/AEM.03130-18.

48. Schwab C, Mastrangelo M, Corsetti A, Gänzle M. 2008. Formation ofoligosaccharides and polysaccharides by Lactobacillus reuteri LTH5448and Weissella cibaria 10M in sorghum sourdoughs. Cereal Chem 85:679 – 684. https://doi.org/10.1094/CCHEM-85-5-0679.

49. Zhao X, Gänzle MG. 2018. Genetic and phenotypic analysis of carbohy-drate metabolism and transport in Lactobacillus reuteri. Int J Food Mi-crobiol 272:12–21. https://doi.org/10.1016/j.ijfoodmicro.2018.02.021.

50. Bankevich A, Nurk S, Antipov D, Gurevich AA, Dvorkin M, Kulikov AS,Lesin VM, Nikolenko SI, Pham S, Prjibelski AD, Pyshkin AV, Sirotkin AV,Vyahhi N, Tesler G, Alekseyev MA, Pevzner PA. 2012. SPAdes: a newgenome assembly algorithm and its applications to single-cell se-quencing. J Comput Biol 19:455– 477. https://doi.org/10.1089/cmb.2012.0021.

51. Bosi E, Donati B, Galardini M, Brunetti S, Sagot M-F, Lió P, Crescenzi P,Fani R, Fondi M. 2015. MeDuSa: a multi-draft based scaffolder. Bioinfor-matics 31:2443–2451. https://doi.org/10.1093/bioinformatics/btv171.

52. Aziz RK, Bartels D, Best AA, DeJongh M, Disz T, Edwards RA, Formsma K,Gerdes S, Glass EM, Kubal M, Meyer F, Olsen GJ, Olson R, Osterman AL,Overbeek RA, McNeil LK, Paarmann D, Paczian T, Parrello B, Pusch GD,Reich C, Stevens R, Vassieva O, Vonstein V, Wilke A, Zagnitko O. 2008.The RAST server: rapid annotations using subsystems technology. BMCGenomics 9:75. https://doi.org/10.1186/1471-2164-9-75.

53. Mitchell A, Sangrador-Vegas A, Quinn AF, McAnulla C, Binns D, Nuka G,McWilliam H, Chang H-Y, Maslen J, Fraser M, Scheremetjew M, Jones P,Lopez R, Hunter S, Pesseat S, Yong S-Y, Li W. 2014. InterProScan 5:genome-scale protein function classification. Bioinformatics 30:1236 –1240. https://doi.org/10.1093/bioinformatics/btu031.

54. Mitchell AL, Sangrador-Vegas A, Luciani A, Madeira F, Nuka G, SalazarGA, Chang H-Y, Richardson LJ, Qureshi MA, Fraser MI, Blum M, RawlingsND, Lopez R, El-Gebali S, Pesseat S, Yong S-Y, Potter SC, Paysan-LafosseT, Finn RD, Marchler-Bauer A, Thanki N, Mi H, Thomas PD, Natale DA,Tosatto SCE, Necci M, Orengo C, Sillitoe I, Attwood TK, Babbitt PC, BrownSD, Bork P, Bridge A, Rivoire C, Sigrist CJA, Redaschi N, Pandurangan AP,Gough J, Haft DR, Sutton GG, Huang H, Letunic I. 2019. InterPro in 2019:improving coverage, classification and access to protein sequence an-notations. Nucleic Acids Res 47:D351–D360. https://doi.org/10.1093/nar/gky1100.

55. Pfaffl MW. 2001. A new mathematical model for relative quantification inreal-time RT–PCR. Nucleic Acids Res 29:e45. https://doi.org/10.1093/nar/29.9.e45.

56. Teixeira JS, Abdi R, Su MSW, Schwab C, Gänzle MG. 2013. Functionalcharacterization of sucrose phosphorylase and scrR, a regulator of su-crose metabolism in Lactobacillus reuteri. Food Microbiol 36:432– 439.https://doi.org/10.1016/j.fm.2013.07.011.

57. Zhang S, Oh J-H, Alexander LM, Özçam M, van Pijkeren J-P, Zhang S, OhJ-H, Alexander LM, Özçam M, van Pijkeren J-P. 2018. D-Alanyl-D-alanineligase as a broad-host-range counterselection marker in vancomycin-




http://aem.asm

.org/D

ownloaded from

https://doi.org/10.1128/AEM.02413-14https://doi.org/10.1016/j.fm.2018.08.009https://doi.org/10.1016/j.fm.2018.08.009https://doi.org/10.1128/AEM.02194-16https://doi.org/10.1128/AEM.02194-16https://doi.org/10.1016/j.cofs.2015.03.001https://doi.org/10.1128/AEM.02116-15https://doi.org/10.1128/AEM.02116-15https://doi.org/10.1006/jsbi.1999.4139https://doi.org/10.1002/mnfr.201200384https://doi.org/10.1128/AEM.04272-13https://doi.org/10.1128/AEM.04272-13https://doi.org/10.1007/s10295-010-0709-6https://doi.org/10.1128/aem.66.8.3368-3375.2000https://doi.org/10.1128/aem.66.8.3368-3375.2000https://doi.org/10.1093/femsre/fux030https://doi.org/10.1016/j.fm.2011.05.001https://doi.org/10.1128/aem.66.12.5322-5328.2000https://doi.org/10.1128/aem.66.12.5322-5328.2000https://doi.org/10.1021/jf803016phttps://doi.org/10.1021/jf803016phttps://doi.org/10.1007/s00253-016-7414-yhttps://doi.org/10.1038/ismej.2011.161https://doi.org/10.1038/s41396-018-0211-0https://doi.org/10.1128/aem.68.2.673-690.2002https://doi.org/10.1080/19490976.2015.1047129https://doi.org/10.1080/19490976.2015.1047129https://doi.org/10.1128/aem.66.9.3966-3973.2000https://doi.org/10.1128/aem.66.9.3966-3973.2000https://doi.org/10.1128/AEM.00668-06https://doi.org/10.1099/ijs.0.63311-0https://doi.org/10.1128/AEM.03130-18https://doi.org/10.1128/AEM.03130-18https://doi.org/10.1094/CCHEM-85-5-0679https://doi.org/10.1016/j.ijfoodmicro.2018.02.021https://doi.org/10.1089/cmb.2012.0021https://doi.org/10.1089/cmb.2012.0021https://doi.org/10.1093/bioinformatics/btv171https://doi.org/10.1186/1471-2164-9-75https://doi.org/10.1093/bioinformatics/btu031https://doi.org/10.1093/nar/gky1100https://doi.org/10.1093/nar/gky1100https://doi.org/10.1093/nar/29.9.e45https://doi.org/10.1093/nar/29.9.e45https://doi.org/10.1016/j.fm.2013.07.011https://aem.asm.orghttp://aem.asm.org/

resistant lactic acid bacteria. J Bacteriol 200:e00607-17. https://doi.org/10.1128/JB.00607-17.

58. de Kok S, Stanton LH, Slaby T, Durot M, Holmes VF, Patel KG, Platt D,Shapland EB, Serber Z, Dean J, Newman JD, Chandran SS. 2014. Rapidand reliable DNA assembly via ligase cycling reaction. ACS Synth Biol3:97–106. https://doi.org/10.1021/sb4001992.

59. Leenhouts K, Buist G, Bolhuis A, Ten Berge A, Kiel J, Mierau I, DabrowskaM, Venema G, Kok J. 1996. A general system for generating unlabelledgene replacements in bacterial chromosomes. Mol Gen Genet 253:217–224. https://doi.org/10.1007/s004380050315.

60. Sanders JW, Venema G, Kok J, Leenhouts K. 1998. Identification of asodium chloride-regulated promoter in Lactococcus lactis by single-copy chromosomal fusion with a reporter gene. Mol Gen Genet 257:681– 685. https://doi.org/10.1007/s004380050697.

61. Edgar RC. 2004. MUSCLE: a multiple sequence alignment method with

reduced time and space complexity. BMC Bioinformatics 5:113. https://doi.org/10.1186/1471-2105-5-113.

62. Trifinopoulos J, Nguyen L-T, Minh BQ, von Haeseler A. 2016. W-IQ-TREE:a fast online phylogenetic tool for maximum likelihood analysis. NucleicAcids Res 44:W232–W235. https://doi.org/10.1093/nar/gkw256.

63. Kalyaanamoorthy S, Minh BQ, Wong TKF, von Haeseler A, Jermiin LS.2017. ModelFinder: fast model selection for accurate phylogeneticestimates. Nat Methods 14:587–589. https://doi.org/10.1038/nmeth.4285.

64. Hoang DT, Chernomor O, von Haeseler A, Minh BQ, Vinh LS. 2018.UFBoot2: improving the ultrafast bootstrap approximation. Mol Biol Evol35:518 –522. https://doi.org/10.1093/molbev/msx281.

65. Letunic I, Bork P. 2016. Interactive Tree of Life (iTOL) v3: an online toolfor the display and annotation of phylogenetic and other trees. NucleicAcids Res 44:W242–W245. https://doi.org/10.1093/nar/gkw290.




http://aem.asm

.org/D

ownloaded from

https://doi.org/10.1128/JB.00607-17https://doi.org/10.1128/JB.00607-17https://doi.org/10.1021/sb4001992https://doi.org/10.1007/s004380050315https://doi.org/10.1007/s004380050697https://doi.org/10.1186/1471-2105-5-113https://doi.org/10.1186/1471-2105-5-113https://doi.org/10.1093/nar/gkw256https://doi.org/10.1038/nmeth.4285https://doi.org/10.1038/nmeth.4285https://doi.org/10.1093/molbev/msx281https://doi.org/10.1093/nar/gkw290https://aem.asm.orghttp://aem.asm.org/

RESULTSHydroxycinnamic acid metabolism of heterofermentative lactobacilli. Identification of putative phenolic acid reductases. Expression profile of putative phenolic acid reductases. Phylogenetic analysis of major genes involved in hydroxycinnamic acid metabolism. Comparison of genotype and phenotype in lactobacilli.

DISCUSSIONMATERIALS AND METHODSBacterial strains and growth conditions. Chemicals. Hydroxycinnamic acid metabolism of lactic aid bacterial strains. Comparative genomics and sequence analysis. RNA isolation and quantification of relative gene expression by RT-qPCR. Construction of L. rossiae par1 and par2 mutants. Phylogenetic analysis. Data availability.

ACKNOWLEDGMENTSREFERENCES

Documents

Genetic Determinants of Hydroxycinnamic Acid Metabolism in ...the Lactobacillus species observed in this study, genotype and phenotype are com-pared in Table 2. In lactobacilli, genotype