crystallization communications\(\def\hfill{\hskip 5em}\def\hfil{\hskip 3em}\def\eqno#1{\hfil {#1}}\)

Journal logoSTRUCTURAL BIOLOGY
COMMUNICATIONS
ISSN: 2053-230X

Preliminary crystallographic studies of glucose de­hydrogenase from the promiscuous Entner–Doudoroff pathway in the hyperthermophilic archaeon Sulfolobus solfataricus

aCentre for Biomolecular Sciences, University of St Andrews, North Haugh, St Andrews, Fife KY16 9ST, Scotland, and bCentre for Extremophile Research, Department of Biology and Biochemistry, University of Bath, Claverton Down, Bath BA2 7AY, England
*Correspondence e-mail: glt2@st-andrews.ac.uk

(Received 20 October 2004; accepted 25 November 2004; online 24 December 2004)

The hyperthermophilic archaeon Sulfolobus solfataricus grows optimally above 353 K and can metabolize glucose and its C4 epimer galactose via a non-phosphorylative variant of the Entner–Doudoroff pathway involving catalytically promiscuous enzymes that can operate with both sugars. The initial oxidation step is catalysed by glucose dehydrogenase (SsGDH), which can utilize both NAD and NADP as cofactors. The enzyme operates with glucose and galactose at similar catalytic efficiency, while its substrate profile also includes a range of other five- and six-carbon sugars. Crystals of the 164 kDa SsGDH homotetramer have been grown under a variety of conditions. The best crystals to date diffract to 1.8 Å on a synchrotron source, have orthorhombic symmetry and belong to space group P21212. Attempts are being made to solve the structure by MAD and MR.

1. Introduction

The hyperthermophilic archaeon Sulfolobus solfataricus grows optimally at 353–358 K and pH 2–4 utilizing a wide range of carbon and energy sources (Grogan, 1989[Grogan, D. W. (1989). J. Bacteriol. 171, 6710-6719.]). Glucose metabolism in this organism proceeds via a non-phosphorylative variant of the classical Entner–Doudoroff pathway, in which one molecule of glucose produces two molecules of pyruvate with no net yield of ATP (De Rosa et al., 1984[De Rosa, M., Gambaccorta, A., Nicolaus, B., Giardina, P., Poerio, E. & Buonocore, V. (1984). Biochem. J. 224, 407-414.]). Analogous pathways have been detected in the thermoacidophilic archaea Sulfolobus acidocaldarius (Selig et al., 1997[Selig, M., Xavier, K. B., Santos, H. & Schönheit, P. (1997). Arch. Microbiol. 167, 217-232.]), Thermoplasma acidophilum (Budgen & Danson, 1986[Budgen, N. & Danson, M. J. (1986). FEBS Lett. 196, 207-210.]) and Thermoproteus tenax (Siebers & Hensel, 1993[Siebers, B. & Hensel, R. (1993). FEMS Microbiol. Lett. 111, 1-8.]), as well as strains of Aspergillus fungi (Elzainy et al., 1973[Elzainy, T. A., Hassan, M. M. & Allam, A. M. (1973). J. Bacteriol. 114, 457-459.]; Elshafei, 1989[Elshafei, A. M. (1989). Acta Biotechnol. 9, 485-489.]).

In S. solfataricus, the enzymes of this non-phosphorylative Entner–Doudoroff pathway have been shown to possess substrate promiscuity (Lamble et al., 2003[Lamble, H. J., Heyer, N. I., Bull, S. D., Hough, D. W. & Danson, M. J. (2003). J. Biol. Chem. 278, 34066-34072.]) and are therefore able to metabolize both glucose and its C4 epimer galactose. The first reaction step involves the NAD(P)-dependent oxidation of glucose to gluconate, or galactose to galactonate, catalysed by glucose dehydrogenase (GDH; EC 1.1.1.47). Gluconate and galactonate are subsequently dehydrated by gluconate dehydratase (GD) to give 2-keto-3-deoxygluconate (KDG) and 2-keto-3-deoxygalactonate (KDGal), respectively (Lamble et al., 2004[Lamble, H. J., Milburn, C. C., Taylor, G. L., Hough, D. W. & Danson, M. J. (2004). FEBS Lett. 576, 133-136.]). KDG aldolase (KDGA) then catalyses the aldolate cleavage of KDG or KDGal, both giving rise to glyceraldehyde and pyruvate (Lamble et al., 2003[Lamble, H. J., Heyer, N. I., Bull, S. D., Hough, D. W. & Danson, M. J. (2003). J. Biol. Chem. 278, 34066-34072.]). The glyceraldehyde molecule is converted to a second molecule of pyruvate by a further four catalytic steps.

The substrate promiscuity of these enzymes may have physio­logical significance and when considered alongside other evidence (Grogan, 1991[Grogan, D. W. (1991). Appl. Environ. Microbiol. 57, 1644-1649.]; Cusdin et al., 1996[Cusdin, F. S., Robinson, M. J., Holman, G. D., Hough, D. W. & Danson, M. J. (1996). FEBS Lett. 387, 193-195.]; Albers et al., 1999[Albers, S.-V., Elferink, M. G. L., Charlebois, R. L., Sensen, C. W., Driessen, A. J. M. & Konings, W. N. (1999). J. Bacteriol. 181, 4285-4291.]) suggests that the organism at no point distinguishes between glucose and galactose during uptake and catabolism. S. solfataricus is considered to be an opportunistic heterotroph (Schafer, 1996[Schafer, G. (1996). Biochim. Biophys. Acta, 1277, 163-200.]) and the pathway promiscuity observed in its sugar metabolism could be an adaptation to the hostile thermoacidophilic environment of the organism, allowing it to scavenge efficiently for energy substrates. This metabolic promiscuity may also reflect an early evolutionary state and is likely to extend to thermoacidophilic archaea closely related to S. solfataricus, such as Thermoplasma acidophilum. This is supported by studies of T. acidophilum GDH (TaGDH), which also has a high activity with glucose and galactose (Smith et al., 1989[Smith, L. D., Budgen, N., Bungard, S. J., Danson, M. J. & Hough, D. W. (1989). Biochem. J. 261, 973-977.]).

Kinetic studies of recombinant S. solfataricus GDH (SsGDH) have shown it to have indistinguishable properties from enzyme purified from the organism. Dual cofactor specificity for NAD and NADP has been demonstrated and catalytic activity has been measured with a number of five- and six-carbon sugar substrates, in addition to D-­glucose and D-galactose for which the enzyme displays a high catalytic efficiency. The substrate profile indicates specificity for the glucose-specific stereo-configuration only at the C2 and C3 carbon positions, while there is no apparent preference for the configuration at the other positions (Lamble et al., 2003[Lamble, H. J., Heyer, N. I., Bull, S. D., Hough, D. W. & Danson, M. J. (2003). J. Biol. Chem. 278, 34066-34072.]). The involvement of a metal ion (most probably Zn2+) in enzyme catalysis is supported by a 60% reduction in activity observed in the presence of EDTA (Lamble et al., 2003[Lamble, H. J., Heyer, N. I., Bull, S. D., Hough, D. W. & Danson, M. J. (2003). J. Biol. Chem. 278, 34066-34072.]).

The SsGDH monomer is composed of 366 amino-acid residues with a molecular weight of 41 kDa and forms tetrameric assemblies in solution (Lamble et al., 2003[Lamble, H. J., Heyer, N. I., Bull, S. D., Hough, D. W. & Danson, M. J. (2003). J. Biol. Chem. 278, 34066-34072.]). The amino-acid sequence has shown that the enzyme is a putative member of the medium-chain alcohol/polyol dehydrogenase/reductase branch of the superfamily of pyridine-nucleotide-dependent alcohol/polyol/sugar dehydrogenases (Edwards et al., 1996[Edwards, K. J., Barton, J. D., Rossjohn, J., Thorn, J. M., Taylor, G. L. & Ollis, D. L. (1996). Arch. Biochem. Biophys. 328, 173-183.]). This family of enzymes includes TaGDH, horse liver alcohol dehydrogenase (LADH), yeast alcohol dehydrogenase (YADH) and sheep liver sorbitol dehydrogenase (LSDH). These enzymes are characterized by a chain length of 350–375 residues and conserved zinc-binding and nucleotide-binding sites. Preference for either NAD or NADP is determined by variations within a GXGXXG/A fingerprint motif and associated charged residues at structurally conserved positions (Wierenga et al., 1985[Wierenga, R. K., De Maeyer, M. C. H. & Hol, W. G. J. (1985). Biochemistry, 24, 1346-1357.]). Sequence alignments have shown that SsGDH possesses a GXGXXG motif (residues 188–193), as well as residues Asn211 and Arg213 at structurally conserved positions involved in nucleotide binding in other members of the family. Kinetic characterization revealed dual cofactor specificity and Arg213 is expected to help stabilize the adenosine-2-phosphate of NADP (Lamble et al., 2003[Lamble, H. J., Heyer, N. I., Bull, S. D., Hough, D. W. & Danson, M. J. (2003). J. Biol. Chem. 278, 34066-34072.]). SsGDH also possesses conserved cysteines and histidines equivalent to the residues involved in binding Zn atoms with putative structural and catalytic roles in the family (Lamble et al., 2003[Lamble, H. J., Heyer, N. I., Bull, S. D., Hough, D. W. & Danson, M. J. (2003). J. Biol. Chem. 278, 34066-34072.]).

Knowledge of the SsGDH structure would allow rationalization of its dual cofactor specificity and unusual substrate profile. The characterization of this enzyme from S. solfataricus forms an integral part of the wider study of the organism's variant Entner–Doudoroff pathway, which is currently under way by the authors. The lack of facial selectivity and substrate specificity of SsKDGA is already the subject of detailed structural investigations (Theodossis et al., 2004[Theodossis, A., Walden, H., Westwick, E. J., Connaris, H., Lamble, H. J., Hough, D. W., Danson, M. J. & Taylor G. L. (2004). J. Biol. Chem. 279, 43886-43892.]). The aim of this research is to achieve a comprehensive understanding of substrate promiscuity in these enzymes, with the possibility of manipulating them to effect specific biotransformations. Such potential is greatly improved by the catalysts' thermostability and specificity for non-phosphorylated substrates.

2. Expression and purification

The GDH gene from S. solfataricus was cloned into the expression vector pREC7/NdeI, which was provided by Dr L. C. Kurz (Washington University School of Medicine, St Louis, USA), and transformed into Escherichia coli JM109. Cells were grown in the presence of ampicillin for 21 h at 310 K without induction and then lysed with BugBuster (Novagen). Cell extracts were heat-treated at 353 K for 30 min and precipitated proteins were removed by centrifugation. Initially, SsGDH was purified from the supernatant by gel filtration using a Superdex 200 16-60 column (Pharmacia) equilibrated in 50 mM Tris HCl pH 7.5 containing 20 mM MgCl2 and 0.2 M NaCl (prep-A). SsGDH activity was demonstrated spectrophotometrically by the increase in absorbance at 340 nm corresponding to enzyme-dependent NADP reduction, as described in Lamble et al. (2003[Lamble, H. J., Heyer, N. I., Bull, S. D., Hough, D. W. & Danson, M. J. (2003). J. Biol. Chem. 278, 34066-34072.]). A revised protocol (prep-B) was used to obtain the best diffracting crystals, in which the gel-filtration buffer was altered to 50 mM Tris–HCl pH 7.5, 20 mM MgCl2, 5 mM NaCl to enable further purification. The protein was then dialysed into 50 mM MES pH 5.5 containing 20 mM MgCl2 and was subsequently applied to a 5 ml bed volume of Reactive Red-500 dye affinity media (Sigma-Aldrich). SsGDH was eluted using the MES buffer with a 0–1.5 M NaCl elution gradient, with SsGDH eluting between 0.5 and 0.7 M NaCl and remaining contaminants eluting at lower salt concentrations. Finally, the pure protein was dialysed into the crystallization buffer of 50 mM Tris–HCl pH 7.5, 20 mM MgCl2.

3. Crystallization and data collection

Early crystallization trials using the hanging-drop vapour-diffusion method yielded crystals diffracting to 7 Å in-house when mounted in thin-walled glass capillary tubes. This primary screen identified PEG and propan-2-ol as successful precipitants and a subsequent sparse-matrix screen of these conditions using the sitting-drop method gave rise to a number of hits. Following optimization of the conditions, improved crystals grew in 0.1 M HEPES pH 6.0, with 13%(w/v) PEG 4000, 8%(v/v) propan-2-ol and 20 mM MgCl2 using protein obtained by prep-A at a concentration of 3.5 mg ml−1 (Fig. 1[link]a). All sitting drops were formed by mixing 2 µl protein solution with 2 µl mother liquor. Crystals reached maturity within two weeks and were positively identified as being composed of SsGDH by mass spectrometry. Following cryoprotection by stepwise equilibration against mother liquor containing 10%(v/v) followed by 20%(v/v) glycerol, crystals were tested in-house at 100 K and diffracted to 3.3 Å. The crystals had orthorhombic symmetry, with unit-cell parameters a = 68.6, b = 88.4, c = 138.8 Å. Data were collected, integrated in MOSFLM (Leslie, 1992[Leslie, A. G. W. (1992). Jnt CCP4/ESF-EACBM Newsl. Protein Crystallogr. 26.]) and then scaled/merged in SCALA (Evans, 1997[Evans, P.R. (1997). Proceedings of the CCP4 Study Weekend. Recent Advances In Phasing, edited by K. S. Wilson, G. Davies, A. W. Ashton & S. Bailey, pp. 97-102. Warrington: Daresbury Laboratory.]; Table 1[link]). The observed systematic absences and calculated merging statistics were consistent with space group P21212.

Table 1
Data-collection statistics

Values in parentheses refer to the highest resolution shell.

  In-house ID14-3
Wavelength (A) 1.54178 0.913
Resolution limits (Å) 50.5–3.3 (3.48–3.30) 148.4–1.8 (1.9–1.8)
Space group P21212 P21212
Unit-cell parameters (Å) a = 68.6, b = 88.4, c = 138.8 a = 68.5, b = 90.3, c = 138.9
No. observations 67816 276584
No. unique reflections 13193 77522
Completeness (%) 99.6 (99.2) 96.7 (92.9)
Rmerge (%) 15.5 (30.1) 6.9 (26.7)
I/σ(I)〉 11.1 (5.3) 12.8 (4.0)
Multiplicity 5.2 (4.9) 3.6 (3.5)
Wilson B2) 43.8 18.145
Rmerge = [\textstyle \sum_{hkl}\sum_{i} |I_{hkl,i} - \langle I_{hkl}\rangle|/][\textstyle \sum_{hkl}\langle I_{hkl} \rangle].
[Figure 1]
Figure 1
(a) Orthorhombic crystals of SsGDH belonging to space group P21212, grown from 0.1 M HEPES pH 6, 13%(w/v) PEG 4000, 8%(v/v) propan-2-ol, 20 mM MgCl2 with 3.5 mg ml−1 protein. (b) Optimized crystals diffracting to 1.8 Å; the largest crystal is approximately 0.3 × 0.1 × 0.1 mm in size. The black bars correspond to 0.1 mm.

Further optimization of these crystals was carried out using protein from prep-B, with the best crystals growing to approximately 0.3 × 0.1 × 0.1 mm in size (Fig. 1[link]b) and being formed from a mother liquor containing 12%(w/v) PEG 4000, 0.1 M MES pH 5.75 and 3%(v/v) propan-2-ol. Although lower concentrations of propan-2-ol led to larger crystals, these were later found to be highly anisotropic. These crystals were cryoprotected using the method described for the previous crystals. Data were collected to 1.8 Å at the ESRF (ID14-3) on a MAR CCD detector and processed in the same way to give the statistics shown in Table 1[link].

4. Crystal analysis

The unit-cell volume suggests the presence of a GDH dimer in the asymmetric unit with an estimated solvent content of 53% and a Matthews coefficient of 2.6 Å3 Da−1 (Matthews, 1968[Matthews, B. W. (1968). J. Mol. Biol. 33, 491-497.]). Calibrated gel filtration of SsGDH shows the enzyme to exist as a tetramer in solution (Lamble et al., 2003[Lamble, H. J., Heyer, N. I., Bull, S. D., Hough, D. W. & Danson, M. J. (2003). J. Biol. Chem. 278, 34066-34072.]) and SsGDH is likely to form a 222 tetramer, as seen in TaGDH, formed by operation of the crystallographic twofold axis along c. A self-rotation function calculated using POLARRFN (Collaborative Computational Project, Number 4, 1994[Collaborative Computational Project, Number 4 (1994). Acta Cryst. D50, 760-763.]; Fig. 2[link]) reveals no obvious NCS twofold axes in the ab plane, suggesting that the NCS twofolds are parallel to the crystallographic a or b axes. This is confirmed by calculation of a native Patterson using data from 10 to 4 Å, which shows a peak 27% the height of the origin at (0.5, 0.5, 0.5). This would suggest that the NCS twofolds are along a and b and that the origin of the tetramer lies a quarter of the way along c.

[Figure 2]
Figure 2
Self-rotation function stereographic projection of the κ = 180° section for SsGDH, calculated with ESRF data from 10 to 4 Å resolution and a sphere of 20 Å using the POLARRFN program (Collaborative Computational Project, Number 4, 1994[Collaborative Computational Project, Number 4 (1994). Acta Cryst. D50, 760-763.]). Spherical polar angles are defined as φ, the angle from the Cartesian x axis (a) on the xy plane (ab); ω, the angle from the z axis (c); κ, the rotation around the axis defined by φω.

5. Attempts at molecular replacement

The SWISSPROT database was searched for homologues of SsGDH and it was found that the greatest sequence identity (32%) is shared with TaGDH, while only 20% identity is shared with horse LADH. The latter enzyme was used as a model in map improvement for solving the TaGDH structure (John et al., 1994[John, J., Crennell, S. J., Hough, D. W., Danson, M. J. & Taylor, G. L. (1994). Structure, 2, 385-393.]) and despite sharing only 19% sequence identity, the two proteins were shown to have significant structural homology.

Secondary-structure predictions for SsGDH using PredictProtein (Rost, 1996[Rost, B. (1996). Methods Enzymol. 266, 525-539.]) also suggested strong structural similarities with TaGDH, so that SsGDH is also expected to possess a central nucleotide-binding domain that forms a Rossmann-fold structure as well as a catalytic domain formed from both the amino-terminal and carboxy-terminal regions of the protein. Local alignments identified the highest sequence identity (40%) between the first 183 residues of SsGDH and the first 186 residues of TaGDH, which includes the coordination sites for both the catalytic and structural Zn atoms present in TaGDH.

Efforts are being made towards molecular replacement in AMoRe (Navaza, 1993[Navaza, J. (1993). Acta Cryst. D49, 588-591.]) and CNS (Brünger et al., 1998[Brünger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W., Jiang, J.-S., Kuszewski, J., Nilges, M., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T. & Warren, G. L. (1998). Acta Cryst. D54, 905-921.]) using a variety of TaGDH-derived polyalanine and full residue models, including the full-length chain, Rossmann-fold region and catalytic domain. Multiple-wavelength anomalous dispersion (MAD) phasing using the scattering by the catalytic/structural Zn atoms in the enzyme is also being attempted.

Footnotes

These authors contributed equally to this work.

Acknowledgements

This work was supported by the Engineering and Physical Sciences Research Council, UK and the Biotechnology and Biological Sciences Research Council, UK. Staff at the European Synchrotron Radiation Facility are thanked for their assistance, as is the European Union for funds to access the facility.

References

First citationAlbers, S.-V., Elferink, M. G. L., Charlebois, R. L., Sensen, C. W., Driessen, A. J. M. & Konings, W. N. (1999). J. Bacteriol. 181, 4285–4291.  Web of Science PubMed CAS Google Scholar
First citationBrünger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W., Jiang, J.-S., Kuszewski, J., Nilges, M., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T. & Warren, G. L. (1998). Acta Cryst. D54, 905–921.  Web of Science CrossRef IUCr Journals Google Scholar
First citationBudgen, N. & Danson, M. J. (1986). FEBS Lett. 196, 207–210.  CrossRef CAS Web of Science Google Scholar
First citationCollaborative Computational Project, Number 4 (1994). Acta Cryst. D50, 760–763.  CrossRef IUCr Journals Google Scholar
First citationCusdin, F. S., Robinson, M. J., Holman, G. D., Hough, D. W. & Danson, M. J. (1996). FEBS Lett. 387, 193–195.  CrossRef CAS PubMed Web of Science Google Scholar
First citationDe Rosa, M., Gambaccorta, A., Nicolaus, B., Giardina, P., Poerio, E. & Buonocore, V. (1984). Biochem. J. 224, 407–414.  CAS PubMed Web of Science Google Scholar
First citationEdwards, K. J., Barton, J. D., Rossjohn, J., Thorn, J. M., Taylor, G. L. & Ollis, D. L. (1996). Arch. Biochem. Biophys. 328, 173–183.  CrossRef CAS PubMed Web of Science Google Scholar
First citationElshafei, A. M. (1989). Acta Biotechnol. 9, 485–489.  CrossRef CAS Web of Science Google Scholar
First citationElzainy, T. A., Hassan, M. M. & Allam, A. M. (1973). J. Bacteriol. 114, 457–459.  CAS PubMed Web of Science Google Scholar
First citationEvans, P.R. (1997). Proceedings of the CCP4 Study Weekend. Recent Advances In Phasing, edited by K. S. Wilson, G. Davies, A. W. Ashton & S. Bailey, pp. 97–102. Warrington: Daresbury Laboratory.  Google Scholar
First citationGrogan, D. W. (1989). J. Bacteriol. 171, 6710–6719.  CAS PubMed Web of Science Google Scholar
First citationGrogan, D. W. (1991). Appl. Environ. Microbiol. 57, 1644–1649.  PubMed CAS Google Scholar
First citationJohn, J., Crennell, S. J., Hough, D. W., Danson, M. J. & Taylor, G. L. (1994). Structure, 2, 385–393.  CrossRef CAS PubMed Web of Science Google Scholar
First citationLamble, H. J., Heyer, N. I., Bull, S. D., Hough, D. W. & Danson, M. J. (2003). J. Biol. Chem. 278, 34066–34072.  Web of Science CrossRef PubMed CAS Google Scholar
First citationLamble, H. J., Milburn, C. C., Taylor, G. L., Hough, D. W. & Danson, M. J. (2004). FEBS Lett. 576, 133–136.  Web of Science CrossRef PubMed CAS Google Scholar
First citationLeslie, A. G. W. (1992). Jnt CCP4/ESF–EACBM Newsl. Protein Crystallogr. 26Google Scholar
First citationMatthews, B. W. (1968). J. Mol. Biol. 33, 491–497.  CrossRef CAS PubMed Web of Science Google Scholar
First citationNavaza, J. (1993). Acta Cryst. D49, 588–591.  CrossRef CAS Web of Science IUCr Journals Google Scholar
First citationRost, B. (1996). Methods Enzymol. 266, 525–539.  CrossRef CAS PubMed Web of Science Google Scholar
First citationSchafer, G. (1996). Biochim. Biophys. Acta, 1277, 163–200.  CAS PubMed Web of Science Google Scholar
First citationSelig, M., Xavier, K. B., Santos, H. & Schönheit, P. (1997). Arch. Microbiol. 167, 217–232.  CAS PubMed Web of Science Google Scholar
First citationSiebers, B. & Hensel, R. (1993). FEMS Microbiol. Lett. 111, 1–8.  CrossRef CAS Google Scholar
First citationSmith, L. D., Budgen, N., Bungard, S. J., Danson, M. J. & Hough, D. W. (1989). Biochem. J. 261, 973–977.  CAS PubMed Web of Science Google Scholar
First citationTheodossis, A., Walden, H., Westwick, E. J., Connaris, H., Lamble, H. J., Hough, D. W., Danson, M. J. & Taylor G. L. (2004). J. Biol. Chem. 279, 43886–43892.  Web of Science CrossRef PubMed CAS Google Scholar
First citationWierenga, R. K., De Maeyer, M. C. H. & Hol, W. G. J. (1985). Biochemistry, 24, 1346–1357.  CrossRef CAS Web of Science Google Scholar

© International Union of Crystallography. Prior permission is not required to reproduce short quotations, tables and figures from this article, provided the original authors and source are cited. For more information, click here.

Journal logoSTRUCTURAL BIOLOGY
COMMUNICATIONS
ISSN: 2053-230X
Follow Acta Cryst. F
Sign up for e-alerts
Follow Acta Cryst. on Twitter
Follow us on facebook
Sign up for RSS feeds