A histidine‐rich extension of the mitochondrial F ₀ subunit ATP6 from the ice worm Mesenchytraeus solifugus increases ATP synthase activity in bacteria

Abstract

Bioenergetic profiles of psychrophiles across domains of life are unusual in that intracellular ATP levels increase with declining temperature. Whole‐transcriptome sequencing of the glacier ice worm Mesenchytraeus solifugus revealed a unique C‐terminal extension on the ATP6 protein, which forms part of the proton pore of mitochondrial ATP synthase (Complex V). This extension, positioned near the proton exit pore, comprises alternating histidine residues thought to increase proton flux through Complex V leading to elevated ATP synthesis. To test this hypothesis, we fused the M. solifugus C‐terminal extension to Escherichia coli AtpB (the ATP6 orthologue) and observed a ~ 5‐fold increase in ATP synthesis. This enhancement was unidirectional as we observed no change to ATP hydrolysis rates. These findings offer an avenue for identifying critical factors associated with ice worm adaptation.

The glacier ice worm Mesenchytraeus solifugus survives year‐round at 0 °C. Its ATP6 subunit, which forms a regulatory component of the proton pore in mitochondrial ATP synthase, has a carboxy‐terminal extension not found in any other organism examined to date. Here, we show that fusion of this extension to the homologous AtpB protein in E. coli results in enhanced ATP production.

Abbreviations

ACMA, 9‐amino‐6‐chloro‐2‐methoxyacridine

BCA, bicinchoninic acid

CCCP, carbonyl cyanide m‐chlorophenyl hydrazone

HEPES, 4‐(2‐Hydroxyethyl)piperazine‐1‐ethane‐sulfonic acid

MOPS, 3‐(N‐morpholino)propanesulfonic acid

Pmf, proton motive force

Glacier ice worms—Mesenchytraeus solifugus of North America (Fig. 1A) and Sinenchytraeus glacialis of the Tibetan Plateau—are the only macroscopic organisms known to reside permanently in ice. Although they display close similarities with other segmented worms (Annelida: Clitellata), ice worms differ by their unusual bioenergetic profile. Specifically, intracellular ATP levels are observed to increase with declining temperature [1, 2], which is distinctly opposite to that observed in mesophilic/thermophilic taxa examined to date [2]. This has been interpreted as a compensatory mechanism whereby elevated ATP increases the probability of molecular collisions, thus enabling biochemical reactions to proceed at cold temperatures [2]. The ice worm's unique ability to maintain unusually high ATP levels remains an intriguing unknown, but here we identify a functional domain of the M. solifugus F ₀ ATP6 subunit that increases efficacy of ATP synthesis by ~ 5‐fold when transferred onto the orthologous subunit, AtpB, in bacteria. This domain comprises 18 amino acids with a notable pattern of alternating histidine residues, forming a carboxy‐terminal extension of ATP6 that projects into the mitochondrial matrix [3, 4]. Our data suggest that the carboxy‐terminal domain may function to shuttle protons away from the F ₀ exit pore, thus increasing rotor velocity and consequently ATP production.

Fig. 1.

The carboxy‐terminal extension of the mitochondrial F ₀ ATP6 subunit from Mesenchytraeus solifugus enhances ATP production. (A) Glacier ice worm, M. solifugus, collected from its native environment. (B) Sequence alignment of ATP6 carboxy‐termini from congener Mesenchytraeus species identifies an 18 residue extension in M. solifugus, which displays homology to numerous bacterial ion transporters. (Mped, M. pedatus; Mhyd, M. hydrius; MsByr, M. solifugus Byron Glacier; MsMMt, M. solifugus Mariner Mt. Glacier; Pseud, Pseudomonadota sp., Brev, Brevundimonas sp.). (C) Western blot of E. coli membranes shows the relative expression of AtpB in the indicated strains (N.S. is a nonspecific band used as a loading control). (D) qRT‐PCR shows the relative expression of the indicated ATP synthase subunits in E. coli (n = 3 biological replicates, error bars are SEM, *P < 0.05, two‐tailed t‐test). (E) Addition of the C‐terminal extension to the chromosomal copy of atpB (exAtpB) enables growth on M9‐acetate media. Deletion of atpB restricts growth to media supplemented with fermentable sugars, such as glucose.

Materials and methods

Bacterial strains, plasmids, and growth conditions

The specific strains, plasmids, and primers utilized in this study can be found in Tables S1–S3. Details of the strain construction are available in Data S1. For routine culturing of E. coli wild‐type strain MG1655 and its derivatives, bacteria were cultured at 37 °C in Luria‐Bertani (LB) medium. For growth to ensure ATP production by oxidative phosphorylation, cells were grown in M9 minimal media with 0.2% (w/v) acetate as the carbon source. Acetate is a nonfermentable carbon source which prevents ATP production by fermentation. C. crescentus cells were grown in peptone–yeast extract (PYE) media at 30 °C. When necessary, antibiotics were added at the following concentrations: kanamycin 30 μg·mL⁻¹ in broth and 50 μg·mL⁻¹ in agar for E. coli and 5 μg·mL⁻¹ in broth and 25 μg·mL⁻¹ in agar for C. crescentus; chloramphenicol 20 μg·mL⁻¹ in broth and 30 μg·mL⁻¹ in agar for E. coli, ampicillin 50 μg·mL⁻¹ in broth and 100 μg·mL⁻¹ in agar for E. coli.

Preparation of inverted membrane vesicles

Membrane vesicles were prepared largely as previously described [5]. Briefly, cells were grown to an A600 of 1.5, harvested by centrifugation at 10 000  g for 5 min, and resuspended in 3 mL MOPS buffer (50 mm MOPS, 10 mm MgCl₂, pH 7.0). Cells were lysed by three passages through a French press at 16 000 psi. Unbroken cells were removed by centrifugation at 10 000  g for 10 min, and the membranes were collected from the supernatant by centrifugation at 38 000  g for 1 h. Membrane pellets were resuspended in 3 mL of MOPS buffer containing 10% glycerol, before a second centrifugation at 38 000  g for 1 h. The supernatant was removed, and pellets were resuspended in 200 μL MOPS buffer. Protein concentration was measured using the BCA protein assay kit ThermoFisher Scientific (Carlsbad, CA, USA).

ATP synthase activity assay

Membranes were diluted to a final concentration of 0.5 μg·μL⁻¹ in MOPS buffer (50 mm MOPS, 10 mm MgCl₂, pH 7.5). A solution containing 20 μL of the diluted membrane sample was used in a total reaction volume of 200 μL containing 50 mm MOPS, pH 7.5, 2.5 mm NADH (to establish a proton gradient), 40 μL luciferase enzyme and substrate mix (CLS II ATP Bioluminescence Kit, Roche (Basel, Switzerland)), and the indicated concentrations of ADP and potassium phosphate. ATP production was measured via luminescence on a BMG CLARIOstar plate reader. ATP levels were quantified using a standard curve of ATP. Since the expression of AtpB is decreased in the exatpB strain, ATP synthase activity was normalized to AtpB expression as determined by western blot. Kinetic parameters were fit using originpro (Origin Lab Corporation (Northampton, MA, USA)).

For experiments varying NADH concentration, activity assays were performed using excess ADP and phosphate (0.5 and 3.75 mm, respectively) such that the limiting factor in the reaction was the proton gradient.

For experiments varying temperature, reactions were carried out for 1 min at the indicated temperature before heating the samples to 95 °C to stop enzymatic activity. After cooling, ATP levels were measured using the above luciferase activity. The normalized activity data were transformed to a linear function according to the Arrhenius equation:

$$\ln \left(V\right)=\ln \left(A\right)-\frac{E_a}{R}\frac{1}{T}$$

Q ₁₀ temperature coefficients were determined using the equation:

$$Q_{10}={\left(\frac{V_2}{V_1}\right)}^{\raisebox{1ex}{${10}^{°}\mathrm{C}$}\!\left/ \!\raisebox{-1ex}{$\left(T_2-T_1\right)$}\right.}$$

Western blotting

Proteins were resolved by SDS/PAGE on a 12% acrylamide gel. After transferring proteins to a nitrocellulose membrane, AtpB was detecting using anti‐AtpB at 1 : 1000 (National BioResource Project (Mishima, Japan)). LamB was detected as a loading control using anti‐LamB antiserum at 1 : 30 000 (kind gift from Marcin Grabowicz, Emory University) [6]. Bands were detected using horseradish peroxidase‐conjugated secondary antibodies and enhanced chemiluminescence reagents (Cytiva (Marlborough, MA, USA)) and imaged on a Bio‐Rad (Hercules, CA, USA) Chemidoc MP. Band intensities were quantified using Bio‐Rad image lab (v 6.0.1); a series of exposures was collected for each blot and those with bands flagged by the software as saturated were not used for normalization.

Quantitative RT‐PCR (qRT‐PCR)

RNA was extracted from bacterial cultures using the Qiagen (Hilden, Germany) RNeasy kit. Following DNase digestion, RNA (25 ng·μL⁻¹) was reverse transcribed using the High‐Capacity cDNA Reverse Transcription Kit (Applied Biosystems). cDNA (1 μL) was used as a template in a 10 μL qRT‐PCR reaction performed with Power SYBR reagent (Applied Biosystems (Waltham, MA, USA)). qRT‐PCR was performed on an ABI QuantStudio 6 using the ΔΔC _t method. rpoD was used as the loading control.

ACMA fluorescence assays

Assay buffer was prepared containing 50 mm HEPES, 5 mm MgCl₂, 300 mm KCl, and the pH adjusted to 7.5 prior to the addition of 0.3 μg·mL⁻¹ 9‐amino‐6‐chloro‐2‐methoxyacridine (ACMA). A total of 50 μg of membrane protein was added to assay buffer for a final volume of 175.8 μL. Samples were added to a black 96‐well plate and ATP hydrolysis was initiated by adding 24.2 μL of 16.5 mm ATP (final concentration was 2 mm). Fluorescence loss was monitored over 2 min (excitation 410 nm, emission 490 nm) in a BMG CLARIOstar plate reader. Finally, 1.02 μL of 4.89 mm carbonyl cyanide m‐chlorophenylhydrazone (CCCP) was added and fluorescence recovery was monitored for an additional 2 min.

To monitor ACMA quenching upon NADH addition, ACMA assay buffer was prepared the same as above, with inverted membranes diluted to 0.25 μg·μL⁻¹. 195 μL of this mixture was then added to a black 96‐well plate and fluorescence was measured for 2 min (excitation 410 nm, emission 490 nm) to serve as a background reading. After this equilibration period, 5 μL of NADH stock solution was added to give a final concentration 0.1 mm, and a time course of fluorescence was immediately recorded.

Intracellular ATP measurements

ATP extractions were performed using a modified protocol from [1]. Briefly, bacterial cultures in biological triplicate were grown overnight at 37 °C in LB broth before back‐diluting into 5 mL of fresh media at an A600 of 0.05. Cultures were incubated for 75 min, to reach an A600 ~ 0.2–0.4, before being chilled on ice and final A600 measurements were made. A 1 mL sample of culture was pelleted at 6000  g for 2 min, the supernatant was removed, and the pellet resuspended in 150 μL HEPES buffer (50 mm HEPES, 50 mAU·mL⁻¹ proteinase K). Samples were lysed by incubation at 50 °C for 30 min and afterward kept on ice. A master mix of 140 μL MOPS buffer (50 mm MOPS, 10 mm MgCl₂, pH 7.5) and 40 μL luciferin/luciferase (CLS II ATP Bioluminescence Kit, Roche) per sample per reaction was made. 180 μL of this master mix and 20 μL of sample were added to a white Costar flat‐bottom 96‐well plate and luminescence was immediately measured. Experiments were performed in technical triplicate, and an ATP standard was utilized to determine ATP concentration. ATP concentrations were normalized to the A600 of the individual samples.

Alphafold structure predictions

alphafold (version 2.2) was used to generate monomer structure predictions using the reduced database [7]. alphafold was run on Amarel, the Rutgers University high performance computing cluster.

Results

The carboxy‐terminal extension on M. solifugus ATP6 is not found on other ATP6 protein sequences currently deposited in GenBank (Fig. 1B). Since manipulation of mitochondrially encoded genes is challenging in general, and currently not possible in M. solifugus, we investigated the impact of this extension by engineering a chromosomal knock‐in of a fusion protein comprising the E. coli ATP6 homolog, AtpB, with the 18 residue carboxy‐terminal extension (exAtpB) (Fig. S1A). The extension sequence was inserted in‐frame at the atpB chromosomal locus such that this was the only copy of atpB in the bacterial genome, and its expression was driven by the native promoter. To confirm membrane insertion of exAtpB, western blots were performed on purified membrane fractions which consistently demonstrated lower levels of AtpB in exAtpB relative to wild‐type (Fig. 1C and Fig. S1B). The decrease in AtpB protein was consistent with downregulated expression of the Complex V operon which showed a similar decrease in the expression of atpA, atpB, and atpD (Fig. 1D). Despite the decrease in Complex V expression, wild‐type and exAtpB strains had identical growth rates when grown in M9 minimal media with glucose or acetate as the carbon source (Fig. 1E); acetate is a nonfermentable carbon source and requires oxidative phosphorylation for growth. Taken together, these data demonstrated that ~ 25% less ATP synthase machinery was required to maintain optimal growth in the exAtpB fusion strain.

The reduced expression of the atp operon in the exAtpB strain could be explained by several different mechanisms. First, there may be an epistatic effect of atpB expression on the remainder of the operon as deletion of atpB resulted in lower levels of atpA and atpD expression (Fig. 1D). Alternatively, since E. coli tunes the expression of the ATP synthase operon to optimize growth rate [8], excess ATP produced by exAtpB could lead to negative feedback and downregulation of the ATP synthetic machinery. Indeed, even the effects of atpB deletion may be explained by this negative feedback since inhibition of ATP synthase paradoxically leads to an increase in cellular ATP levels, likely through enhanced glycolytic activity (Fig. S1C) [9].

To directly assess ATP synthase activity, inverted membrane preparations were incubated with NADH, ADP, and phosphate while ATP synthesis was monitored by standard luciferin–luciferase assays. To account for the variation in Complex V expression, ATP synthesis rates were normalized to AtpB expression based on western blot analysis (Fig. S1B). By independently varying ADP and phosphate concentrations, our data showed that the exAtpB carboxy‐terminus fusion protein resulted in a 5‐fold increase in apparent V _max (Fig. 2A,B, and Table 1). Importantly, this increase was observed at ADP and phosphate concentrations consistent with physiological levels in E. coli (0.12 mm ADP [10] and 2–10 mm phosphate [11, 12]).

Fig. 2.

Biochemical characterization of the C‐terminal extension. (A, B) ATP production by E. coli membrane fractions was compared between wild‐type (WT) and exatpB strains. Rates were normalized to AtpB protein expression. Michaelis–Menten kinetic parameters were determined by curve‐fitting (n = 4 biological replicates, error bars are SEM) and are presented in Table 1.

Table 1.

Kinetic parameters of E. coli Complex V.

Strain	V max,app (nmol ATP·s−1·mg−1 mem. Protein)	ADP K m,app (μm)	Phosphate K m,app (μm)
Wild‐type	5.2 ± 0.1	17 ± 2	240 ± 20
exAtpB	26.5 ± 1.8	23 ± 3	320 ± 20

Increased ATP production in exAtpB could be attributed directly to enhanced Complex V activity; alternatively, we considered whether this mutation resulted in feedback to Complex I leading to an increased proton motive force (pmf) which could similarly increase ATP synthesis. Generation of the pH gradient by Complex I was measured via 9‐amino‐6‐chloro‐2‐methoxyacridine (ACMA) quenching as protons are pumped into the vesicle lumen upon incubation with NADH. No difference was observed between WT and exAtpB vesicles (Fig. 3A). Similarly, qRT‐PCR data showed no changes to expression of Complex I subunit genes nuoF and nuoA in the exAtpB strain (Fig. 3B).

Fig. 3.

exAtpB expression does not affect Complex I activity. (A) ACMA quenching was quantified in WT and exAtpB membranes treated with 0.1 mm NADH (n = 4 biological replicates, error bars are SD). (B) qRT‐PCR shows the relative expression of the indicated Complex I subunits in E. coli (n = 3 biological replicates, error bars are SEM).

While no differences in Complex I activity were observed, we considered whether there may be differences in Complex V activities at nonsaturating concentrations of NADH. ATP synthase activity was measured as a function of varying NADH concentration, with higher activity associated with faster proton translocation (Fig. 4A). At each NADH concentration, exAtpB membranes consistently showed higher activity. Interestingly, increased proton flux in exAtpB appears to be unidirectional since no increase of the inward proton flux was observed under conditions of ATP hydrolysis (Fig. 4B).

Fig. 4.

Characterization of exAtpB activity dependence on NADH and temperature. (A) Activity of ATP synthase was measured in the presence of indicated concentrations of NADH, 0.5 mm ADP, and 3.75 mm phosphate, followed by normalization to protein expression. (B) ATP hydrolysis and proton flux were assessed using ACMA fluorescence. ATP was added at t = 0 to initiate proton flux, and the uncoupler carbonyl cyanide m‐chlorophenylhydrazone (CCCP) was added at the indicated time (n = 3 biological replicates, error bars are SEM). (C) Maximum velocity was measured for WT and exAtpB membranes at 10, 16, 22, 30, and 37 °C and normalized to AtpB expression (n = 3 biological replicates, error bars are SEM). The data were transformed into a linear graph based on the Arrhenius equation, and Q ₁₀ coefficients were determined.

Since the C‐terminal extension was derived from an ice worm that survives under permanently cold physiological conditions, we tested whether the C‐terminal extension enhanced ATP synthase activity as a function of declining temperature. The rate of ATP synthesis varied with temperature as predicted by the Arrhenius equation; however, no temperature‐dependent differences were observed between wild‐type and exAtpB samples with both AtpB variants having similar Q ₁₀ temperature coefficients (Fig. 4C).

We hypothesized that the alternating histidine residues may play a functional role in increasing ATP synthesis by increasing proton flux through the F₀ pore, based on the role of similar histidine‐rich domains in prokaryotic ion transporters [13, 14, 15]. To test this model, histidine residues were mutated in an E. coli inducible expression plasmid; however, any attempt to mutate one or more histidines to other amino acids was toxic, leading to nonsense mutations or deletions in the promoter region that inhibited expression. Attempts included mutating histidines within the extension at position 12, 14, or 16 individually or in combination to alanine, as well as H12G and H12Q mutations. We cannot explain this instability other than non‐native variants of this extension appear to be toxic to E. coli. Note also that mitochondrial atp6 is a hypersensitive locus; for example, more than 30 independent single nucleotide polymorphisms in ATP6 have been identified that lead to mitochondrial dysfunction in humans [16].

To determine whether increased exAtpB efficacy in E. coli, a Gammaproteobacterium, was an anomaly, we tested whether this effect would occur in the phylogenetically distant Alphaproteobacterial species, Caulobacter crescentus. Using a native‐locus extension knock‐in mutation, a similar increase in V _max,app was observed (Fig. 5A), despite limited sequence conservation between the two AtpB proteins (19% identical and 39% similar). Additionally, the extension had no effect on the expression of atpB or ccna_00371 (B' subunit), and only a slight decrease in ccna_00372 (C subunit) (Fig. 5B). This may reflect differences in the organization of the ATP synthase genomic loci; E. coli encodes all of Complex V in a single operon while the F ₀ and F ₁ subunits are encoded on separate loci in C. crescentus. Alternatively, the metabolic feedback circuitry may differ as E. coli is capable of fermentation whereas C. crescentus is an obligate aerobe.

Fig. 5.

The C‐terminal extension increases ATP production in C. crescentus. (A) ATP production by C. crescentus membrane fractions was compared between wild‐type (WT) and exatpB strains. Michaelis–Menten kinetic parameters were determined by curve‐fitting (n = 3 biological replicates, error bars are SEM). (B) qRT‐PCR shows the relative expression of the indicated ATP synthase subunits in C. crescentus (n = 3 biological replicates, error bars are SEM, *P < 0.05, two‐tailed t‐test).

Discussion

Collectively, our findings demonstrate an enhanced level of ATP synthase activity conferred by the ice worm ATP6 carboxy‐terminal extension. This is surprising since mitochondrial Complex V is known to operate at very high efficiency [17, 18]. Although this extension is not present on other ATP6 proteins currently deposited in GenBank, similar amino acid sequences are found in a number of microbial ion transporters (Fig. 1B). alphafold‐predicted structures of ice worm ATP6, E. coli exAtpB, and bacterial ion transporters with homologous alternating histidine motifs show that the motif is found in disordered regions located near the transporter pore opening (Fig. S1D). Given the increased rate of ATP synthesis observed in exAtpB, we hypothesize that the C‐terminal extension of ice worm ATP6 may facilitate proton translocation through the Complex V pore, though the mechanism for this remains unclear. An alternative model is that while the pmf is the same in WT and exAtpB cells, the decrease in Complex V expression in exAtpB could lead to a greater pmf : Complex V ratio. We view this model as less likely since pmf is not rate limiting for ATP synthesis [5, 19].

Expression of exatpB from its native promoter leads to decreased levels of AtpB in membrane preparations and transcriptional downregulation of the atp operon. While these data might be directly related, it is also important to consider that the C‐terminal extension may affect ATP synthase assembly and maturation [20], and this modified subunit may not be efficiently incorporated into a functional Complex V. In this case, our values for ATP synthase rates in exAtpB would be underestimations since we cannot distinguish between functional and nonfunctional incorporation of AtpB into Complex V via western blot. If, instead, the lower levels of exAtpB are due to transcriptional downregulation, this could be due to epistasis or negative feedback from enhanced ATP production. Transcription of other genes in the atp operon is inhibited in the ΔatpB strain (Fig. 1D), consistent with an epistatic effect. However, we cannot rule out negative feedback from ATP because cellular levels of ATP are actually higher in ΔatpB (Fig. S1C). Clearly, the mechanism(s) underlying the regulation of Complex V expression requires further study.

The sequence homology to bacterial ion transporters suggests that M. solifugus may have acquired this domain by horizontal gene transfer (HGT) from a glacial bacterium [4]. While there are many examples of intra‐HGT events between mitochondria and the nucleus, reports of HGT between bacteria and mitochondria are absent, thus making the ice worm ATP6 carboxy‐terminal extension particularly noteworthy.

The possibility that the ice worm ATP6 extension independently promotes gains in intracellular ATP at lower temperatures counters the Arrhenius relationship and seems unlikely given that membrane fractions from wild‐type and exAtpB cells show similar Q ₁₀ values for ATP synthesis (Fig. 4C). We note that the kinetics of the electron‐transport chain are also affected by the changes in temperature; however, if the C‐terminal extension profoundly decreased the activation energy of ATP synthesis, we would have expected the slope of the activity graph to be affected. More likely, the influence of the C‐terminal extension on ATP synthesis functions in a steady‐state, intracellular ATP regulatory network, and one must consider the various metabolic reactions that lead to energy stasis. Furthermore, we know that the ATP6 extension is not the only adaptation responsible for psychrophily since other cold‐adapted Mesenchytraeus species, including the snow‐worm M. hydrius, lacks any ATP6 C‐terminal modification (Fig. 1B). Further analysis is required to determine the mechanism by which this protein domain increases ATP synthase activity, but it offers an avenue for identifying critical factors associated with ice worm adaptation, and possibly for enhancing ATP production in other organisms and contexts such as photosynthesis in plant chloroplast or human mitochondrial dysfunction.

Author contributions

TD, DHS, and EAK conceived and designed the project; TD and EAK developed the methodologies; TD acquired the data; TD, DHS, and EAK analyzed and interpreted the data; TD, DHS, and EAK wrote the paper.

Peer review

The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer‐review/10.1002/1873‐3468.15100.

Supporting information

Data S1. Supplemental methods.

Fig. S1. Additional characterization of exAtpB.

Table S1. Strains used in this study.

Table S2. Plasmids used in this study.

Table S3. Primers used in this study.

Data accessibility

The data that support the findings of this study are available in Figs 1, 2, 3, 4, 5, Table 1, and the Supporting Information of this article.