Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat
Sign Up
PreprintCited by 4Protein Structure and DynamicsGenomics and Chromatin DynamicsRNA Research and Splicing

The DOI matches its Crossref record; 1 of the 1 references it cites were found in a registry or an index.

Insight into the structural basis for dual nucleic acid - recognition by the Scaffold Attachment Factor B2 protein

Open access
Korn, Sophie M., von Ehr, Julian, Dhamotharan, Karthikeyan, Tants, Jan-Niklas, Abele, Rupert, Schlundt, Andreas·2022·International Journal of Molecular Sciences
DOI 10.1101/2022.12.28.522114PMC PMC9958909

Preprint — published as 10.3390/ijms24043286 · Paper page

Open this paper in CoChat — free accountOpen PDF

Abstract

The family of scaffold attachment factor B (SAFB) proteins comprises three members and was first identified as binders of the nuclear matrix/scaffold. Over the past two decades, SAFBs were shown to act in DNA repair, mRNA/(l)ncRNA processing, and as part of protein complexes with chromatin-modifying enzymes. SAFB proteins are approximately-100-kDa-sized dual nucleic acid-binding proteins with dedicated domains in an otherwise largely unstructured context, but whether and how they discriminate DNA- and RNA-binding has remained enigmatic. We here provide the SAFB2 DNA- and RNA-binding SAP and RRM domains in their functional boundaries and use solution NMR spectroscopy to ascribe DNA- and RNA-binding functions. We give insight into their target nucleic acid preferences and map the interfaces with respective nucleic acids on sparse data-derived SAP and RRM domain structures. Further, we provide evidence that the SAP domain exhibits intra-domain dynamics and a potential tendency to dimerise, which may expand its specifically targeted DNA sequence range. Our data provide a first molecular basis of and a starting point towards deciphering DNA- and RNA-binding functions of SAFB2 on the molecular level and serve a basis for understanding its localization to specific regions of chromatin and its involvement in the processing of specific RNA species.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1101/2022.12.28.522114

Cites

1 of 1 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 Scaffold attachment factor proteins (SAF) type A (hnRNP U) and B (SAFB1, SAFB2 and SLTM). (A) Domain organizations of human SAFA, SAFB1, SAFB2 and SLTM as annotated in Uniprot [30] (IDs: Q00839, Q15424, Q14151, Q9NWH9). (B) Sequence alignment of the SAP domain (based on domain boundaries of the SAFB2 SAP defined in this study) with residues color-coded by conservation amongst the four proteins according to Clustal Omega [31] (black—no conservation, light orange—weak-to-strong conservation, dark orange—full conservation). Above, secondary structure elements derived from the SCS analysis for SAFB2 SAP in Figure 2 are given, with rectangles indicating α-helices. The helix–loop–helix DNA-binding motif spans helices 2 (H2) and 3 (H3). (C) Sequence alignment of the SAFB RRM domains (boundaries are based on those of the SAFB2 RRM domain defined in this study) with residues colored according to the degree of conservation as by Clustal Omega (black—no conservation, light green—weak-to-strong conservation, dark green—full conservation). Above, the two conserved RNA-binding motifs (RNP) and secondary structure elements derived from the SCS analysis for SAFB2 RRM in Figure 3 are given, with rectangles indicating α-helices and arrows indicating β-strands, respectively.
Figure 1 Scaffold attachment factor proteins (SAF) type A (hnRNP U) and B (SAFB1, SAFB2 and SLTM). (A) Domain organizations of human SAFA, SAFB1, SAFB2 and SLTM as annotated in Uniprot [30] (IDs: Q00839, Q15424, Q14151, Q9NWH9). (B) Sequence alignment of the SAP domain (based on domain boundaries of the SAFB2 SAP defined in this study) with residues color-coded by conservation amongst the four proteins according to Clustal Omega [31] (black—no conservation, light orange—weak-to-strong conservation, dark orange—full conservation). Above, secondary structure elements derived from the SCS analysis for SAFB2 SAP in Figure 2 are given, with rectangles indicating α-helices. The helix–loop–helix DNA-binding motif spans helices 2 (H2) and 3 (H3). (C) Sequence alignment of the SAFB RRM domains (boundaries are based on those of the SAFB2 RRM domain defined in this study) with residues colored according to the degree of conservation as by Clustal Omega (black—no conservation, light green—weak-to-strong conservation, dark green—full conservation). Above, the two conserved RNA-binding motifs (RNP) and secondary structure elements derived from the SCS analysis for SAFB2 RRM in Figure 3 are given, with rectangles indicating α-helices and arrows indicating β-strands, respectively.CC BY · PubMed Central
Figure 2 Resonance assignment of the SAFB2 SAP domain. (A) 1H-15N HSQC of the SAP domain amide groups. Straight lines indicate side chain assignments. (B) Combined Cα/Cβ carbon secondary chemical shifts (SCS) (upper panel, as suggested by [32]—rectangles indicating α-helices) and the {1H}15N heteronuclear NOE values (lower panel) of SAFB2 SAP plotted against the residue number. Errors are calculated from the program CCPNMR Analysis.
Figure 2 Resonance assignment of the SAFB2 SAP domain. (A) 1H-15N HSQC of the SAP domain amide groups. Straight lines indicate side chain assignments. (B) Combined Cα/Cβ carbon secondary chemical shifts (SCS) (upper panel, as suggested by [32]—rectangles indicating α-helices) and the {1H}15N heteronuclear NOE values (lower panel) of SAFB2 SAP plotted against the residue number. Errors are calculated from the program CCPNMR Analysis.CC BY · PubMed Central
Figure 3 Resonance assignment of the SAFB2 RRM domain. (A) 1H-15N HSQC of the RRM domain amide groups. Assignments of peaks in the red dashed-lined box are shown in the inset zoom-in. See also Supplementary Figure S1B. (B) Combined Cα/Cβ carbon secondary chemical shifts (SCS) (upper panel, as suggested by [32]—rectangles indicating α helices, arrows indicating β-strands) and the {1H}15N heteronuclear NOE values (lower panel) of the SAFB2 RRM plotted against residue numbers. Errors are calculated from the program CCPNMR Analysis.
Figure 3 Resonance assignment of the SAFB2 RRM domain. (A) 1H-15N HSQC of the RRM domain amide groups. Assignments of peaks in the red dashed-lined box are shown in the inset zoom-in. See also Supplementary Figure S1B. (B) Combined Cα/Cβ carbon secondary chemical shifts (SCS) (upper panel, as suggested by [32]—rectangles indicating α helices, arrows indicating β-strands) and the {1H}15N heteronuclear NOE values (lower panel) of the SAFB2 RRM plotted against residue numbers. Errors are calculated from the program CCPNMR Analysis.CC BY · PubMed Central
Figure 4 RNA-binding preferences of the SAFB2 RRM domain. (A) 1H-15N HSQC zoom-ins showing overlays of 2× molar excess of RNAs as indicated (color code depicted in upper left corner) with the apo RRM, respectively. Assignments for RNP residues involved in RNA interaction are included. Addition of G7 resulted in severe precipitation and was, thus, not evaluated accordingly. See Supplementary Figure S4 for full spectra. (B) CSP plots of individual RNAs over amino acid sequence. Indicated above is the RRM secondary structure with RNP1 and RNP2 highlighted. Dotted grey lines represent the CSP average. (C) Mapping of the ten significantly perturbed CS for residues shown as spheres (N-atoms) to a structural model of the SAFB2 RRM domain, i.e., the lowest energy representative of a CS-Rosetta-derived ensemble (Supplementary Figure S2). Maximum CSPs locate to the two RNP motifs and the C-terminal tail, respectively.
Figure 4 RNA-binding preferences of the SAFB2 RRM domain. (A) 1H-15N HSQC zoom-ins showing overlays of 2× molar excess of RNAs as indicated (color code depicted in upper left corner) with the apo RRM, respectively. Assignments for RNP residues involved in RNA interaction are included. Addition of G7 resulted in severe precipitation and was, thus, not evaluated accordingly. See Supplementary Figure S4 for full spectra. (B) CSP plots of individual RNAs over amino acid sequence. Indicated above is the RRM secondary structure with RNP1 and RNP2 highlighted. Dotted grey lines represent the CSP average. (C) Mapping of the ten significantly perturbed CS for residues shown as spheres (N-atoms) to a structural model of the SAFB2 RRM domain, i.e., the lowest energy representative of a CS-Rosetta-derived ensemble (Supplementary Figure S2). Maximum CSPs locate to the two RNP motifs and the C-terminal tail, respectively.CC BY · PubMed Central
Figure 5 DNA-binding preferences of the SAFB2 SAP domain. (A) EMSAs of SAP domain with ss (left) and ds (right) DNAs, either AT-rich (top) or GC-rich (bottom) (with sense sequences as follows: AT13mer 5′-GCAATAAATACG-3′; GC13mer 5′-GCCCCCGCGCCCG-3′). Increasing SAP protein concentration (0–40 µM) is indicated by black bar above gel pictures. EMSAs shown here are representatives of technical triplicates (Supplementary Figure S6). (B) 1H-15N HSQC of SAP titrated with increasing molar ratios (0.5×, 1×, 1.85×, 4×, 6×) of the dsDNA AT16mer (5′-GCGCACAATATAACGC-3′). The most perturbed resonances are labelled with their assignments. The zoom-in of the boxed region below shows the highly conserved Lys53 (see Figure 1), which is most strongly affected. Note the different contour levels used to visualize this peak relative to the settings of the full spectrum above. (C) CSP plots of 6× molar excess of AT16mer plotted over amino acid sequence. Indicated above is the SAP secondary structure. Dotted grey line represents the CSP average, the solid black line the threshold of 2× standard deviation. (D) Mapping of the four significantly perturbed CS for residues shown as spheres (N-atoms) plotted on a structural model of the SAFB2 SAP domain. The structural model is the lowest energy representative of a CS-Rosetta-derived ensemble.
Figure 5 DNA-binding preferences of the SAFB2 SAP domain. (A) EMSAs of SAP domain with ss (left) and ds (right) DNAs, either AT-rich (top) or GC-rich (bottom) (with sense sequences as follows: AT13mer 5′-GCAATAAATACG-3′; GC13mer 5′-GCCCCCGCGCCCG-3′). Increasing SAP protein concentration (0–40 µM) is indicated by black bar above gel pictures. EMSAs shown here are representatives of technical triplicates (Supplementary Figure S6). (B) 1H-15N HSQC of SAP titrated with increasing molar ratios (0.5×, 1×, 1.85×, 4×, 6×) of the dsDNA AT16mer (5′-GCGCACAATATAACGC-3′). The most perturbed resonances are labelled with their assignments. The zoom-in of the boxed region below shows the highly conserved Lys53 (see Figure 1), which is most strongly affected. Note the different contour levels used to visualize this peak relative to the settings of the full spectrum above. (C) CSP plots of 6× molar excess of AT16mer plotted over amino acid sequence. Indicated above is the SAP secondary structure. Dotted grey line represents the CSP average, the solid black line the threshold of 2× standard deviation. (D) Mapping of the four significantly perturbed CS for residues shown as spheres (N-atoms) plotted on a structural model of the SAFB2 SAP domain. The structural model is the lowest energy representative of a CS-Rosetta-derived ensemble.CC BY · PubMed Central
Figure 6 DNA-binding capacity of SAFB2 SAP domain mutants. 1H-15N HSQC overlay of apo SAP proteins (yellow) with 6- or 8-fold excess (red) of AT9mer (5′-CAATATAAC-3′), as indicated. Top panel: SAPWT (left) and SAPR45A (right); bottom panel: SAPS31A (left) and SAPK53A/S54A (right). Contour levels are scaled individually between proteins, for peaks to be visible. Contour levels for each apo and DNA-bound spectra of one protein are adjusted.
Figure 6 DNA-binding capacity of SAFB2 SAP domain mutants. 1H-15N HSQC overlay of apo SAP proteins (yellow) with 6- or 8-fold excess (red) of AT9mer (5′-CAATATAAC-3′), as indicated. Top panel: SAPWT (left) and SAPR45A (right); bottom panel: SAPS31A (left) and SAPK53A/S54A (right). Contour levels are scaled individually between proteins, for peaks to be visible. Contour levels for each apo and DNA-bound spectra of one protein are adjusted.CC BY · PubMed Central
Figure 7 SAFB2 SAP domain mutants. (A) aSEC of SAP wild-type domain and mutants. (B) Estimation of molecular weights (MW) based on MW standards. Shown is the partition coefficient (Kav) plotted as a function of MW. Colored points (in dotted circle) are the Kav from aSEC over their theoretical MWs (expecting a monomer), respectively. Colored points on the fit line are the MWs determined based on the calibration standards (see also Supplementary Figure S12). (C) SEC-MALS-derived MWs of the SAP wild-type domain and mutants compared to the theoretical MW of a mono-mer.
Figure 7 SAFB2 SAP domain mutants. (A) aSEC of SAP wild-type domain and mutants. (B) Estimation of molecular weights (MW) based on MW standards. Shown is the partition coefficient (Kav) plotted as a function of MW. Colored points (in dotted circle) are the Kav from aSEC over their theoretical MWs (expecting a monomer), respectively. Colored points on the fit line are the MWs determined based on the calibration standards (see also Supplementary Figure S12). (C) SEC-MALS-derived MWs of the SAP wild-type domain and mutants compared to the theoretical MW of a mono-mer.CC BY · PubMed Central
Figure 8 Summary of nucleic acid-binding preferences of SAFB2 SAP and RRM domains as determined in this study. In addition, possible inter- and intra-molecular domain interactions based on current experimental knowledge are depicted. SAFB2 and its domains are not shown in scale.
Figure 8 Summary of nucleic acid-binding preferences of SAFB2 SAP and RRM domains as determined in this study. In addition, possible inter- and intra-molecular domain interactions based on current experimental knowledge are depicted. SAFB2 and its domains are not shown in scale.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 1 references

Research works printed but not linked: 82

  1. 10.1016/j.cell.2019.06.001 · DOI · Paper page
  2. 10.1038/nrm3884 · DOI · Paper page
  3. Davis, B.N. ; Hilyard, A.C. ; Nguyen, P.H. ; Lagna, G. ; Hata, A. SMAD proteins bind a conserved RNA sequence to promote microRNA maturation by Drosha. Mol. Cell 2010, 39. · DOI · Paper page
  4. Shi, L. ; Godfrey, W.R. ; Lin, J. ; Zhao, G. ; Kao, P.N. NF90 regulates inducible IL-2 gene expression in T cells. J. Exp. Med. 2007, 204. · DOI
  5. Shim, J. ; Lim, H. ; Yates Iii, J. ; Karin, M. Nuclear export of NF90 is required for interleukin-2 mRNA stabilization. Mol. Cell 2002, 10. · DOI
  6. Kuo, P.H. ; Doudeva, L.G. ; Wang, Y.T. ; Shen, C.K.J. ; Yuan, H.S. Structural insights into TDP43 in nucleic-acid binding and domain interactions. Nucleic Acids Res. 2009, 37. · DOI · Paper page
  7. Lukavsky, P.J. Molecular basis of UG-rich RNA recognition by the human splicing factor TDP43. Nature Struct. Mol. Biol. 2013, 20. · DOI · Paper page
  8. Huang, D.B. Crystal structure of NF-κB (p50)2 complexed to a high-affinity RNA aptamer. Proc. Natl Acad. Sci. USA 2003, 100. · DOI
  9. Müller, C.W. ; Rey, F.A. ; Sodeoka, M. ; Verdine, G.L. ; Harrison, S.C. Structure of the NF-κB p50 homodimer bound to DNA. Nature 1995, 373. · DOI · Paper page
  10. 10.1002/j.1460-2075.1992.tb05331.x · DOI · Paper page
  11. 10.1002/j.1460-2075.1992.tb05422.x · DOI · Paper page
  12. 10.1093/nar/24.5.843 · DOI
  13. 10.1242/jcs.112.10.1465 · DOI
  14. 10.3390/ijms22094982 · DOI
  15. 10.1111/febs.14141 · DOI
  16. 10.1042/BCJ20160649 · DOI
  17. 10.1016/j.molcel.2019.10.001 · DOI
  18. 10.1093/nar/gkt285 · DOI
  19. 10.1042/BST20120030 · DOI
  20. 10.1007/s00395-021-00854-y · DOI
  21. 10.1073/pnas.2005712117 · DOI
  22. 2021, Scaffold attachment factor B1 regulates androgen degradation pathways in prostate cancer, Am J Clin Exp Urol, 9, 337
  23. 10.1016/j.molcel.2020.05.011 · DOI
  24. 10.1038/s41422-021-00515-8 · DOI
  25. 10.1002/jcb.10685 · DOI
  26. 10.1002/jcb.24182 · DOI
  27. 10.1016/j.febslet.2010.11.054 · DOI
  28. 10.1016/j.molcel.2020.01.026 · DOI
  29. 10.15252/embj.2019104154 · DOI
  30. 10.17989/ENCSR558RBK · DOI
  31. 10.1016/j.molcel.2014.04.016 · DOI · Paper page
  32. 10.1016/j.sbi.2008.04.002 · DOI · Paper page
  33. Maris, C. ; Dominguez, C. ; Allain, F.H.T. The RNA recognition motif, a plastic RNA-binding platform to regulate post-transcriptional gene expression. FEBS J. 2005, 272. · DOI · Paper page
  34. 10.17989/ENCSR072VUO · DOI
  35. 10.1093/nar/gkaa1100 · DOI · Paper page
  36. 10.1038/msb.2011.75 · DOI · Paper page
  37. 10.1016/S0968-0004(99)01537-6 · DOI · Paper page
  38. 10.1016/j.jbc.2022.102268 · DOI
  39. 10.1007/BF00211777 · DOI · Paper page
  40. 10.1007/s12104-020-09971-6 · DOI
  41. 10.1007/s12104-020-09995-y · DOI
  42. 10.1021/bi00213a010 · DOI
  43. 10.1021/bi300539t · DOI
  44. 10.1016/j.pnmrs.2018.03.002 · DOI
  45. 10.1073/pnas.1203013109 · DOI
  46. 10.1093/nar/gkq366 · DOI · Paper page
  47. 10.1002/prot.21973 · DOI
  48. 10.1073/pnas.1616745113 · DOI
  49. 10.1006/jmbi.2001.4592 · DOI
  50. 10.1017/S0033583515000207 · DOI
  51. 10.1038/s41467-019-09769-8 · DOI
  52. 10.1093/nar/gkac212 · DOI
  53. 10.1093/nar/gkm615 · DOI
  54. 10.1074/jbc.M105238200 · DOI
  55. 10.1074/jbc.M212988200 · DOI
  56. 10.1242/jcs.03344 · DOI
  57. 10.1016/j.redox.2022.102545 · DOI
  58. 10.1016/j.sbi.2012.11.006 · DOI
  59. 10.1016/j.pnmrs.2014.05.003 · DOI
  60. 10.1093/nar/gkv1303 · DOI
  61. 10.1038/s41467-022-33552-x · DOI
  62. 10.1073/pnas.1303445110 · DOI
  63. 10.1261/rna.065649.118 · DOI
  64. 10.26508/lsa.202000825 · DOI
  65. 10.1038/380646a0 · DOI
  66. 10.1016/j.str.2011.05.002 · DOI
  67. 10.1016/j.str.2021.05.001 · DOI
  68. 10.1074/jbc.RA120.015824 · DOI
  69. 10.1038/nature06152 · DOI
  70. 10.1093/nar/28.1.235 · DOI · Paper page
  71. 10.1002/wrna.1107 · DOI
  72. 10.1093/bib/bbab195 · DOI
  73. 10.1038/s41467-022-30183-0 · DOI
  74. 10.1186/1472-6750-9-61 · DOI · Paper page
  75. 10.1093/nar/gkaa1013 · DOI
  76. 10.7554/eLife.01456 · DOI · Paper page
  77. 1994, Chemical-Shifts as a Tool for Structure Determination, Nuclear Magnetic Resonance, Pt C; Methods in Enzymology, 239, 363, 10.1016/S0076-6879(94)39014-2 · DOI · Paper page
  78. 10.1016/0022-2364(87)90269-1 · DOI
  79. 2005, The CCPN data model for NMR spectroscopy: Development of a software pipeline, Proteins-Structure Function and Bioinformatics, 59, 687, 10.1002/prot.20449 · DOI · Paper page
  80. 10.1093/bioinformatics/btu830 · DOI · Paper page
  81. 10.1073/pnas.0800256105 · DOI · Paper page
  82. 10.1371/journal.pone.0077245 · DOI

Metadata

Authors
Korn, S. M., von Ehr, J., Dhamotharan, K., Tants, J.-N., Abele, R., & Schlundt, A.
Affiliations
Goethe University Frankfurt (DE)
Corresponding author
Schlundt, Andreas
ORCID
Venue
International Journal of Molecular Sciences
Publisher
openRxiv
Published
2022-12-29
Volume
24
Issue
4
Pages
3286-3286
ISSN
1422-0067 (ISSN-L)
DOI
10.1101/2022.12.28.522114
PMC
PMC9958909
OpenAlex
W4319600471
Cited by
4
References
1
Citations by year
2026: 1 · 2025: 1 · 2024: 2
FWCI
0.36
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
  • International Journal of Molecular Sciences (publishedVersion, cc-by)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by-nc)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Publication Server of Goethe University Frankfurt am Main (Goethe University Frankfurt) (submittedVersion, cc-by-nc)
  • Publication Server of Goethe University Frankfurt am Main (Goethe University Frankfurt) (submittedVersion, cc-by)
Funders
  • Johanna Quandt Young Academy (2019/AS01)
  • Deutsche Forschungsgemeinschaft (2019/AS01, SCHL2062/2-1)
  • University of California, San Diego
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Protein Structure and Dynamics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Chromatin Dynamics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › RNA Research and Splicing
Keywords
nucleic acid recognition, DNA binding, RNA binding, SAP domain, RRM domain, domain structure, protein dimerization, chromatin, nuclear matrix, RNA processing, DNA repair
MeSH terms
Base Sequence, Binding Sites, Chromatin, Magnetic Resonance Spectroscopy, RNA, RNA, Messenger

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Insight into the structural basis for dual nucleic acid - recognition by the Scaffold Attachment Factor B2 protein. https://app.cochat.ai/papers/10.1101/2022.12.28.522114
The paper
Korn, S. M., von Ehr, J., Dhamotharan, K., Tants, J.-N., Abele, R., & Schlundt, A. (2022). Insight into the structural basis for dual nucleic acid - recognition by the Scaffold Attachment Factor B2 protein. International Journal of Molecular Sciences. https://doi.org/10.1101/2022.12.28.522114

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.