Identification of a Tyrosine Decarboxylase Gene (tdcA) in Streptococcus thermophilus 1TT45 and Analysis of Its Expression and Tyramine Production in Milk

Abstract

In this study, a tyrosine decarboxylase gene (tdcA) was identified in 1 among 83 Streptococcus thermophilus strains tested. Its sequence, nearly identical to that of a tdcA of Lactobacillus curvatus, indicated a horizontal gene transfer event. Transcription in milk and the formation of critical levels of tyramine were observed in the presence of tyrosine.

Biogenic amine (BA) formation in fermented food has recently been included by the European Food Safety Authority (EFSA) among the safety concerns requiring evaluation. Among these compounds, tyramine is the most common in ripened cheeses and fermented sausages, where it usually accumulates during ripening and can be present in toxicologically relevant levels (>100 mg kg⁻¹) (4). The symptoms of tyramine poisoning, also known as “cheese reaction,” are hypertension due to peripheral vasoconstriction caused by the release of noradrenaline, pupil dilation, and increased respiration frequency and blood sugar level (22).

Tyramine formation is the result of bacterial decarboxylation of tyrosine that is carried out by several microorganisms, including strains of lactic acid bacteria (LAB) belonging to different genera and species (7, 23). Among LAB, Streptococcus thermophilus is of major importance for the dairy industry, where it is commonly used in starter cultures for the manufacture of yogurt, fermented milks, and many varieties of cheese. Recently, a few strains of dairy origin assigned to the species S. thermophilus have been described for their tyraminogenic activity by both physiological assays (15) and consensus PCR on tyrosine decarboxylase genes (tdcA genes) (3). However, the identity of the genetic determinant involved was not unveiled.

The aims of this study were (i) to screen the tyramine production of S. thermophilus in synthetic medium, (ii) to identify the gene responsible for this activity, (iii) to quantify its expression by quantitative real-time reverse transcription-PCR (RT-qPCR), and (iv) to monitor the levels of tyramine production in the presence of physicochemical parameters common to a cheese-making process in order to obtain indications on how to limit the formation of this BA.

Screening of S. thermophilus strains for tyramine production.

Eighty-three S. thermophilus dairy isolates, identified and genotypically characterized in previous studies (12, 20), were investigated. All strains were subcultured in LM17 medium, i.e., M17 medium (Fluka, Italy) supplemented with 0.5% (wt/vol) lactose, at 37°C for 24 h. Initially, the strains were screened for tyramine production according to the method of Bover-Cid and Holzapfel (2). Only the strain S. thermophilus 1TT45 (16S rRNA gene sequence has EMBL/GenBank/DDBJ accession no. FR725449), isolated from Taleggio cheese, gave a positive reaction in plates. Consensus tdcA-targeted PCR with the degenerate primers DEC5/DEC3 (23) on genomic DNA extracted with the method of Rossi et al. (21) from 2-ml cultures generated an amplicon of the expected size of close to 336 bp only from this strain (data not shown). This DNA fragment was purified with the Wizard SV gel and PCR clean-up system (Promega, Italy) and sequenced with primers DEC5/DEC3 by BMR Genomics (Italy). BLAST alignment at NCBI (www.ncbi.nlm.nih.gov) revealed 100% identity with a tdcA gene identified in Lactobacillus curvatus HSCC1737 (EMBL/GenBank/DDBJ accession no. AB086652).

Sequencing of the S. thermophilus 1TT45 tdcA gene.

Sequencing of the entire genetic determinant, aimed at its specific identification and at designing an RT-qPCR test for studying its expression, was undertaken. Indeed, making available a qPCR assay directed to this particular gene and different from the already-described tdcA-specific PCR-based detection methods would permit monitoring of its expression in mixed bacterial ecosystems.

To understand whether the unknown portions of the S. thermophilus tdcA gene were also similar to the corresponding gene in L. curvatus, the specific primers TDECLC-F3 and TDECLC-R3 (Table 1) were designed on the L. curvatus HSCC1737 tdcA gene extremities and paired with DEC5 or DEC3 to obtain two gene regions constituting the entire open reading frame (ORF). The PCR mixture (25 μl) contained 1.5 mM MgCl₂, 1 μM each primer, 100 μM each deoxynucleoside triphosphate (dNTP), 0.025 U μl⁻¹ of GoTaq DNA polymerase (Promega), and approximately 50 ng of genomic DNA. The PCR program comprised initial denaturation at 94°C for 5 min; 30 cycles of 94°C for 30 s, 50°C for 60 s (5 min for the primer pair DEC5/TDECLC-R3), 72°C for 60 s; and final extension at 72°C for 7 min. Amplicons of the expected length were obtained (data not shown). Their sequencing permitted the determination of the entire tdcA ORF of S. thermophilus 1TT45, with uncertainty only in the primer annealing sites. The translated tdcA sequence of S. thermophilus 1TT45, analyzed at http://smart.embl-heidelberg.de/, contained the consensus pattern of pyridoxal phosphate-dependent decarboxylases.

TABLE 1.

Targets and sequences of the primer pairs used in this study

Primera	Sequence (5′→3′)	Target	Function	Nucleotide position	Product size (bp)	Reference or source
DEC5	CGTTGTTGGTGTTGTTGGCACNACNGARGARG	tdcA	Screening for tdcA	869-901b	336	19
DEC3	CCGCCAGCAGAATATGGAAYRTANCCCAT			1205-1176b
TDECLC-F3	ATGAGTAACACTAGTTTTAGTG	tdcA	tdcA amplification, paired with DEC3	1-22b	1,205	This study
TDECLC-R3	TTTACGAAGATCGTAAATAA	tdcA	tdcA amplification, paired with DEC5	1846-1866b	997	This study
STTDEC-1F	CGTTCACAATCAGTTCCTT	tdcA	RT-qPCR	247-266c	166	This study
STTDEC-1R	CCAACTTCTTCTTCCATTTG			413-393c
RPOST-F	ACTGTCATTGTTGCTTGGAATG	rpoA	RT-qPCR	521-542d	114	This study
RPOST-R	AGCTGAGGTTACTGCTGGAGAT			613-634d

^aPrimers DEC5/DEC3, STTDEC-1F/STTDEC-1R, and RPOST-F/RPOST-R were used in pairs for the functions indicated.

^bPosition in the L. curvatus HSCC1737 tdcA sequence (EMBL/GenBank/DDBJ accession no. AB086652).

^cPosition in the S. thermophilus 1TT45 tdcA sequence (EMBL/GenBank/DDBJ accession no. FR682467).

^dPosition in the S. thermophilus LMD-9 rpoA sequence (EMBL/GenBank/DDBJ accession no. CP000419, nucleotide positions 1755506 to 1756444).

BLAST analysis revealed nucleotide sequence identities of 95% to 99% to different L. curvatus tdcA sequences and of approximately 72% to the tdcA genes from Enterococcus spp. and Lactobacillus brevis. The highest identity (99.9%) was shared with the 1,866-bp tdcA sequence of L. curvatus HSCC1737: only two nucleotide substitutions, 443A→C and 606A→G, with a missense mutation Lys→Thr (K148T) for the substitution at position 443, were found. Interestingly, this result indicated a recent horizontal gene transfer (HGT) event between taxonomically unrelated species. HGT of another tdcA gene was previously hypothesized, but only at the intraspecies level (7). Thus, the HGT among different species put in evidence in this study deserves further attention through screening for the presence of tdcA in other LAB sharing the same ecological niches as S. thermophilus and L. curvatus. These are mainly milk, dairy products, sausages, and plant materials.

Determination of the tyraminogenic potential of S. thermophilus 1TT45.

The tyraminogenic potential of S. thermophilus 1TT45 was preliminarily quantified by ion-pair high-pressure liquid chromatography and postcolumn derivatization with ortho-phthalaldehyde according to the method of Hernández-Jover et al. (11), inoculating an initial bacterial number of 5 × 10⁵ CFU ml⁻¹ into LM17 medium and into M17 medium with a lower lactose content (0.1%, wt/vol), with or without the addition of 0.1% (wt/vol) tyrosine (a concentration that can be found in cheese) or 2% (wt/vol) NaCl at 37°C. The latter condition was taken into account in order to obtain indications about the possible influence that salt could exert on tdcA expression and tyramine production in milk. The experiments were carried out in triplicate. The results are shown in Fig. 1, where tyramine production with reference to the growth phases is reported. The strain appeared to be a weak/moderate producer, approaching a tyramine concentration of 100 mg liter⁻¹, when tyrosine was not added. Indeed, the availability of small amounts of precursor in the broth cultures was ensured by the presence of sources of free amino acids. However, when tyrosine was added, the strain behaved as a moderate producer, since the amounts of tyramine reached 400 to 500 mg liter⁻¹ (1). The highest and fastest production took place in the presence of NaCl, indicating positive regulation by osmotic shock, which is undoubtedly the main impact of salt. This effect could be a consequence of the indirect involvement of the tyrosine-decarboxylating pathway in supplying the energy necessary for osmoregulation. Similarly, during growth with a lower lactose concentration (0.1%, wt/vol), tyramine was produced in larger amounts, underlining the role of tyrosine decarboxylation as a secondary energetic route (16). In this trial, statistically significant differences between tyramine production levels were observed in the presence and absence of NaCl (P < 0.05) and with higher and lower lactose concentrations (P < 0.01), as determined by the Student t test. In LM17 medium and in M17 medium plus 0.1% (wt/vol) lactose, the production rate was higher at the beginning of the stationary phase of growth, probably activated by lactose exhaustion, while in the presence of NaCl, the maximum rate was observed during the exponential phase of growth, probably in response to osmotic stress.

FIG. 1.

Growth (dotted lines) and tyramine production (bars) by S. thermophilus 1TT45 at 37°C in LM17 medium and M17 medium supplemented with lactose (0.1%, wt/vol) with or without added tyrosine (0.1%, wt/vol) and NaCl (2%, wt/vol). Error bars show standard deviations.

Study of S. thermophilus 1TT45 tdcA expression in growth conditions common to cheese making.

The qPCR test designed here for the relative quantification of the S. thermophilus tdcA transcript used as a reference the constitutive gene rpoA, encoding the alpha subunit of the RNA polymerase. The latter was reported to show stable expression under different conditions (19). The qPCR primers (Table 1) were designed with the Primer Express version 2.0 software (Applied Biosystems, CA). Primer specificity was confirmed by using BLAST at NCBI. The qPCR mixtures (20 μl), using Platinum SYBR green supermix UDG with ROX (5-carboxy-X-rhodamine) dye (Invitrogen, Italy), contained 0.14 μM each primer and 2 μl of template DNA or cDNA. The PCR program was comprised of initial denaturation at 95°C for 5 min and 40 cycles at 95°C for 30 s and 50°C for 1 min, followed by dissociation of the amplicons in the temperature range of 60 to 95°C.

The calibration curves, constructed by using as targets duplicate 10-fold serial dilutions of the STTDEC-1F/STTDEC-1R and RPOST-F/RPOST-R amplicons and reported in Fig. 2, showed that the qPCR tests developed in this study fulfilled the requirements of recently published guidelines on the use of RT-qPCR (5). The copy number of the target fragments per reaction mixture was adjusted by measuring the optical density at 260 nm (OD₂₆₀) with a BioPhotometer (Eppendorf, Italy). The efficiencies [E = 10^(−1/slope) − 1] of the tdcA- and rpoA-specific qPCR tests, calculated from the respective calibration curves, were 0.81 and 0.9, respectively, and the dissociation curves showed excellent specificity for both amplification reactions (data not shown). The two qPCR tests had different linearity ranges (Fig. 2) and limits of detection (LOD) (2 × 10² and 20 target copies per reaction mixture for tdcA and rpoA, respectively). Similar to the results reported by Ritz et al. (19), who indicated rpoA as a good calibrator, absolute quantification of the transcripts of this gene obtained from fixed amounts of RNA (1 μg) showed a low value for the maximum fold variation (2^−ΔCT = 1.95), at least under the growth conditions used in the following experiments (data not shown).

FIG. 2.

Calibration curves for the tdcA (a) and rpoA (b) qPCR tests.

The tdcA expression level and the production of tyramine were evaluated under different conditions during 7 days of incubation in skim milk (Oxoid, Italy) inoculated with an initial number of 5 × 10⁵ CFU ml⁻¹ of S. thermophilus strain 1TT45. For some cultures, the inoculum was heat treated in a water bath (67°C for 10 s) to simulate a mild thermization process before cheese making. The heat treatment was chosen on the basis of what is done in the technology of some traditional Italian cheeses, for which a very short thermization is carried out to lower the microbial load of raw milk without destroying useful natural microbial components. The cultures were kept at 37°C for 24 h and then transferred to 20°C, an environmental temperature used in cheese ripening. In some samples, tyrosine was added at a concentration of 0.1% (wt/vol) after 8 h to simulate the formation of the free amino acid after proteolysis. NaCl (2%, wt/vol) was then added after 24 h of incubation, as is done in cheese salting. The growth and pH evolution of S. thermophilus 1TT45 exhibited similar trends in all the cultures, reaching maximum values of between 10⁹ and 3 × 10⁹ CFU ml⁻¹ after 24 h and decreasing after day 4 to 2 × 10⁸ to 4 × 10⁸ CFU ml⁻¹ at day 7. The pH fell to 4.27 in 24 h and then decreased slowly to 4.21 at day 7, being slightly higher, by 0.04 units, in the presence of tyrosine.

RNA was extracted from duplicate skim milk cultures, at the different times reported in Fig. 3, as follows: 1 volume of 26.8 mM EDTA, pH 12.0 (18), was added to 4 ml of the milk culture, and the mixture was centrifuged and washed twice with sterile water. The extraction was carried out as described by Torriani et al. (23). RNA integrity, concentration, and purity were checked by electrophoresis on 2% (wt/vol) agarose gel and by measurement of the OD₂₆₀/OD₂₈₀ ratio. One microgram of total RNA was added to 12 μl of RT reaction mixture in 1× RT buffer (50 mM Tris, pH 8.3, 75 mM KCl, 3 mM MgCl₂) containing 10 mM dithiothreitol and 300 μM dNTPs. The RNA was first denatured at 65°C for 5 min. Then, 100 μM random hexamers (Promega), 3 mM MgCl₂, 1 U μl⁻¹ RNasin Plus RNase inhibitor (Promega), and 5 U μl⁻¹ Moloney murine leukemia virus reverse transcriptase (Promega) were added, and the reaction was carried out for 1 h at 37°C.

FIG. 3.

Transcription ratios of the tdcA gene, with rpoA as the reference gene and the 8-h culture as the control, and levels of tyramine produced in skim milk (only the time of culture is indicated), with a heat-treated inoculum (ht), in the presence of 0.1% (wt/vol) tyrosine (tyr), and with 2% (wt/vol) NaCl (salt). The incubation temperatures were 37°C until 24 h and then 20°C.

The tdcA expression ratios, with rpoA expression as the reference, were determined by RT-qPCR and the equation of Pfaffl (17). The latter describes a mathematical model for relative quantification in RT-qPCR that takes into account the reaction efficiency of the PCR tests and the cycle threshold (C_T) values for a gene to be quantified and for a reference gene in a treated sample compared to a control. In this case, the control sample was represented by the 8-h cultures. The two genes were amplified in separate sets of reactions performed in duplicate on the same cDNA samples. The results are reported in Fig. 3.

The basal transcription level of tdcA, according to the C_T values and corresponding copy numbers deduced from the calibration curve in Fig. 2, was at least 2 × 10³ copy numbers per reaction. The transcription ratios increased drastically at the seventh day, especially in the cultures without tyrosine and salt (Fig. 3). This result is in line with the function of decarboxylases in providing metabolic energy when the main carbohydrate sources are depleted and is supported by the fact that all the tested cultures were declining in number at 7 days. The well-known role of TdcA decarboxylases in resistance to acidity (10, 16) was less evident in this trial, since a drastic increase of expression as soon as the pH fell to low levels was not observed.

What is unusual, and is part of the novelty of the results, is that the increase of expression was retarded even in the presence of the substrate. Thus, it can be supposed that a combination of stress conditions, more than single factors, can activate the expression of the S. thermophilus tdcA gene. Notably, the presence of tyrosine did not appear to be involved in this effect, considering that in its absence, the expression ratio was even higher (Fig. 3). The lack of tyramine formation when tyrosine was not supplied is explained by the incapability of S. thermophilus 1TT45 of forming the substrate for the TdcA decarboxylase from milk proteins. However, when tyrosine was added, the levels of BA formed were high, though the expression level was lower than in the absence of the amino acid, indicating a good efficiency of the TdcA enzymatic reaction. The lower enhancement of expression found at 7 days in the case of tyrosine addition could be due to a relief of the harsh growth conditions determined by the increase of the pH, as well as to having provided an essential amino acid as an extra nutritional source. The lower expression in milk in the presence of salt requires further investigation.

The tdcA transcript was undetectable in the cultures inoculated with the heat-treated bacteria. This finding is at this moment unexplainable due to the lack of data in the literature and deserves further investigation aimed at understanding the regulation mechanisms involved and how these can be exploited to prevent tyramine formation by S. thermophilus in dairy products. The lack of expression observed in this study can probably be explicated by investigations on the global changes occurring after heat shock and/or by in-depth studies on the regulation of tdcA. To this aim, additional work will be dedicated to determining the regulatory mechanisms acting on the tdcA gene locus.

In the skim milk cultures, tyramine accumulated at high levels only when tyrosine was supplied (Fig. 3), showing that the ability of S. thermophilus 1TT45 to form this BA depends on the availability of the precursor in the culture medium, as observed in other bacteria by Marcobal et al. (14). This result is explained by the incapability of S. thermophilus of releasing peptides and free amino acids from milk proteins when grown in pure culture. Indeed, only a small number of S. thermophilus strains possess a functional cell wall-associated proteinase (8). In this study, the absence of a proteinase in S. thermophilus 1TT45 was confirmed by analyzing the growth curve of the strain in milk. This was not biphasic, which is different from what is observed in the presence of a proteinase (data not shown) according to Letort et al. (13). Nevertheless, the presence in cheeses of highly proteolytic LAB species would likely allow tyramine formation by S. thermophilus. Though the duration of the trial carried out in this study, i.e., 7 days, was much shorter than a cheese-ripening process, it highlighted that S. thermophilus can accumulate tyramine in the earlier phases of cheese making, so this BA can remain at high levels unless other bacteria degrade it in the prosecution of the ripening process. Since S. thermophilus 1TT45 was able to produce up to 370 mg liter⁻¹ of tyramine in milk containing tyrosine, the accumulation of toxic levels of tyramine in a cheese made with S. thermophilus tyraminogenic strains is possible. Indeed, 10 to 80 mg of tyramine are considered toxicologically relevant doses (6), and these amounts could be consumed by eating a moderate quantity of cheese, e.g., a 100-g portion, containing the concentrations formed in milk by S. thermophilus strains with a tyraminogenic potential similar to that observed in this study.

Some discussion regarding yogurt production is opportune, though the use in this product of well-characterized S. thermophilus cultures can allow the exclusion of tyraminogenic strains from the technological process. Based on our experiment using milk, we could hypothesize that no production of tyramine is expected during yogurt manufacture, since tdcA expression in skim milk without salt increased after an incubation time longer than the duration of the fermentation process for this product. The effect of the higher temperature (42°C) used for yogurt fermentation might not be relevant, since it is an optimal value for this bacterial species and does not imply particular physiological stresses. Tyramine production during storage at refrigeration temperatures is expected to be lower than was found at 20°C in this study, since the rate of the TdcA reaction would probably be slowed.

This is the first study in which a tyraminogenic S. thermophilus strain was investigated in detail by identifying and entirely sequencing the tdcA genetic determinant and following its expression under conditions common to cheese making. Importantly, it emerged that the presence of this gene in S. thermophilus was a consequence of an HGT event between distantly related bacterial species, which is a further risk factor identified in the present investigation.

According to studies indicating that the activities of LAB TdcAs are modulated by growth and environmental conditions (9), our experiments demonstrated that the presence of the tdcA gene, induced mainly under conditions of low energy availability, represents a risk when tyrosine is present in the medium, due to the activity of proteolytic bacteria during cheese ripening.

Further research on tdcA regulation aimed at better defining the risk related to the tyraminogenic potential of S. thermophilus will suggest strategies to prevent tyramine formation and accumulation in dairy products and contribute to the formulation of safety recommendations in this respect. More specifically, the Qualified Presumption of Safety (QPS) status of this species assigned by EFSA should be amended accordingly. A countermeasure like the addition of competing cultures to limit the growth of adventitious LAB carrying this negative physiological trait in traditional cheeses made from raw milk could reduce the tyramine content of these food products. This point deserves particular attention since experimentation in this field is still lacking.

Nucleotide sequence accession number.

The tdcA gene sequence of S. thermophilus 1TT45 has been deposited in the EMBL/GenBank/DDBJ databases under accession no. FR682467.